Skip to main content

Full text of "Sino-Iranica; Chinese contributions to the history of civilization in ancient Iran, with special reference to the history of cultivated plants and products"

See other formats


"LI B R.AR.Y' 



v. 15 

cop. d. 

Return this book on or before the 
Latest Date ■ stamped below. 

Theft, mutilation, and underlining of books 
are reasons for/disciplinary action and may 
result in dismissal from the University. 
Universify of Illinois Library 


D mm 


Htf* HMO 


E Ivi A* '<■ v 1 9^8 






L161— O-1096 




Volume XV, No. 3 



' Field Museum of Natural History 
Publication 201 
Anthropological Series Vol. XV, No. 3 


Chinese Contributions to the History of Civilization 
in Ancient Iran 

With Special Reference to the History of 
Cultivated Plants and Products 


Berthold Laufer 

Curator of Anthropology 

The Blackstone Expedition 





Field Museum of Natural History 

Publication 201 

Anthropological Series Vol. XV, No. 3 


Chinese Contributions to the History of Civilization 
in Ancient Iran 

With Special Reference to the History of 
Cultivated Plants and Products 


Berthold Laufer 

Curator of Anthropology 

The Blackstone Expedition 








Introduction 185 

Sino-Iranica 208 

Alfalfa 208 

The Grapf-Vine 220 

The Pistachio 246 

The Walnut 254 

The Pomegranate 276 

Sesame and Flax \ 288 

The Coriander 297 

The Cucumber 300 

Chive, Onion, and Shallot 302 

Garden Pea and Broad Bean 305 

Saffron and Turmeric 309 

Safflower 324 

Jasmine 329 

Henna 334 

The Balsam-Poplar 339 

Manna 343 

Asafoetida 353 

Galbanum 363 

Oak-Galls 367 

Indigo 370 

Rice 372 

Pepper 374 

Sugar 376 

Myrobalan ...- 378 

The "Gold Peach" 379 

Fu-tse 379 

Brassica 380 

Cummin 383 

The Date-Palm 385 

The Spinach 392 

Sugar Beet and Lettuce 399 

Ricinus 403 

The Almond 405 

The Fig 410 

The Olive 415 


iv Contents, 

\{ Page 

Cassia Pods and Carob 420 

Narcissus 427 

The Balm of Gilead 429 

Note on the Language of Fu-lin 435 

The Water-Melon 438 

Fenugreek 446 

Nux-Vomica 448 

The Carrot 451 

Aromatics 455 

Spikenard, p. 455. — Storax, p. 456.— Myrrh, p. 460. — Putchuck, p. 462. — Styrax 
benjoin, p. 464. 

The Malayan Po-se and Its Products 468 

Alum, p. 474. — Lac, p. 475. — Camphor, p. 478. — Aloes, p. 480. — Amomum, p. 481. — 
P, o-lo-te, p. 482. — Psoralea, p. 483.— Ebony, p. 485. 

Persian Textiles 488 

Brocades, p. 488. — Rugs, p. 492. — Yue no, p. 493. — Woolen §tuSs, p. 496. — Asbestos, 
p. 498. 

Iranian Minerals, Metals, and Precious Stones . . . 503 

Borax, p. 503. — Sal Ammoniac, p. 503. — Litharge, p. 508. — Gold, p. 509. — Oxides 
of Copper, p. 510. — Colored Salt, p. 511. — Zinc, p. 511.— -Steel, p. 515. — 
Se-se, p. 516. — Emerald, p. 518. — Turquois, p. 519. — Lapis Lazuli, p. 520. — 
Diamond, p. 521. — Amber, p. 521. — Coral, p. 523. — Bezoar, p. 525. 

Titles Of the Sasanian Government 529 

Irano-Sinica 535 

The Square Bamboo, p. 535. — Silk, p. 537. — Peach and Apricot, p. 539. — Cinnamon, 
p. 541. — Zedoary, p. 544. — Ginger, p. 545. — Mamiran, p. 546. — Rhubarb, p. 547. — 
Salsola, p. 551. — Emblic Myrobalan, p. 551. — Althaea, p. 551. — Rose of China, 
p. 551.— Mango, p. 552. — Sandal, p. 552. — Birch, p. 552.— Tea, p. 553. — Onyx, 
p. 554. — Tootnague, p. 555. — Saltpetre, p. 555. — Kaolin, p. 556. — Smilax pseudo- 
china, p. 556. — Rag-paper, p. 557. — Paper Money, p. 559. — Chinese Loan-Words 
in Persian, p. 564. — The Chinese in the Alexander Romance, p. 570. 

Appendix I Iranian Elements in Mongol 572 

Appendix II Chinese Elements in Turki 577 

Appendix III The Indian Elements in the Persian Pharma- 
cology of Abu Mansur Muwaffaq . . . 580 

Appendix IV The Basil • • v 5^^ 

Appendix V Additional Notes on Loan-Words in Tibetan 591 

General Index 599 

Botanical Index > 617 

Index of Words 621 



By Berthold Laufer 


If we knew as much about the culture of ancient Iran as about 
ancient Egypt or Babylonia, or even as much as about India or China, 
our notions of cultural developments in Asia would probably be widely 
different from what they are at present. The few literary remains left 
to us in the Old-Persian inscriptions and in the Avesta are insufficient 
to retrace an adequate picture of Iranian life and civilization; and, 
although the records of the classical authors add a few touches here 
and there to this fragment, any attempts at reconstruction, even 
combined with these sources, will remain unsatisfactory. During the 
last decade or so, thanks to a benign dispensation of fate, the Iranian 
horizon has considerably widened: important discoveries made in 
Chinese Turkistan have revealed an abundant literature in two hitherto 
unknown Iranian languages, — the Sogdian and the so-called Eastern 
Iranian. 1 ' We now know that Iranian peoples once covered an immense 
territory, extending all over Chinese Turkistan, migrating into China, 
coming in contact with Chinese, and exerting a profound influence on 
nations of other stock, notably Turks and Chinese. The Iranians were 
the great mediators between the West and the East, conveying the 
heritage of Hellenistic ideas to central and eastern Asia and trans- 
mitting valuable plants and goods of China to the Mediterranean area. 
Their activity is of world-historical significance, but without the 
records of the Chinese we should be unable to grasp the situation 
thoroughly. The Chinese were positive utilitarians and always inter- 
ested in matters of reality: they have bequeathed to us a great amount 
of useful information on Iranian plants, products, animals, minerals, 
customs, and institutions, which is bound to be of great service to 

The following pages represent Chinese contributions to the history 
of civilization in Iran, which aptly fill a lacune in our knowledge of 
Iranian tradition. Chinese records dealing with the history of Iranian 
peoples also contain numerous transcriptions of ancient Iranian words, 

1 Cf., for instance, P. Pelliot, Influences iraniennes en Asie centrale et en 
Extreme-Orient (Paris, 191 1). 


1 86 Sino-Iranica 

part of which have tested the ingenuity of several sinologues and 
historians; but few of these Sino-Iranian terms have been dealt with 
accurately and adequately. While a system for the study of Sino- 
Sanskrit has been successfully established, Sino-Iranian has been 
woefully neglected. The honor of having been the first to apply the 
laws of the phonology of Old Chinese to the study of Sino-Iranica is 
due to Robert Gauthiot. 1 It is to the memory of this great Iranian 
scholar that I wish to dedicate this volume, as a tribute of homage not only 
to the scholar, but no less to the man and hero who gave his life for 
France. 2 Gauthiot was a superior man, a kiiin-tse ^ J" in the sense of 
Confucius, and every line he has written breathes the mind of a thinker 
and a genius. I had long cherished the thought and the hope that I 
might have the privilege of discussing with him the problems treated 
on these pages, which would have considerably gained from his sagacity 
and wide experience — #^A£,^HfrWft^. 

Iranian geographical and tribal names have hitherto been identified 
on historical grounds, some correctly, others inexactly, but an attempt 
to restore the Chinese transcriptions to their correct Iranian prototypes 
has hardly been made. A great amount of hard work remains to be 
, done in this field. 3 In my opinion, it must be our foremost object first 
to record the Chinese transcriptions as exactly as possible in their 
ancient phonetic garb, according to the method so successfully inaugu- 
rated and applied by P. Pelliot and H. Maspero, and then to proceed 
from this secure basis to the reconstruction of the Iranian model. 
The accurate restoration of the Chinese form in accordance with 

1 Cf. his Quelques termes techniques bouddhiques et manich£ens, Journal 
asiatique, 191 1, II, pp. 49-67 (particularly pp. 59 et seq.), and his contributions to 
Chavannes and Pelliot, Traite" manicheen, pp. 27, 42, 58, 132. 

* Gauthiot died on September It, 191 6, at the age of forty, from the effects of a 
wound received as captain of infantry while gallantly leading his company to a 
grand attack, during the first offensive of Artois in the spring of 1915. Cf. the 
obituary notice by A. Meillet in Bull, de la Societe de Linguistique, No. 65, 
pp. 127-132. 

8 I hope to take up this subject in another place, and so give only a few examples 
here. Ta-ho §wi 3^ -|g ^fC is the Ta-ho River on which Su-li, the capital of Persia, 
was situated (Sui Su, Ch. 83, p. 7 b). Hirth (China and the Roman Orient, pp. 198, 
313; also Journal Am. Or. Soc, Vol. XXXIII, 1913, p. 197), by means of a Cantonese 
Tat-hot, has arrived at the identification with the Tigris, adding an Armenian 
Deklath and Pliny's Diglito. Chinese la, however, corresponds neither to ancient 
ti nor de, but only to *tat, dat, dad, dar, d'ar, while ho fisj represents *hat, kat, kad, 
kar, kal. We accordingly have *Dar-kat, or, on the probable assumption that a 
metathesis has taken place, *Dak-rat. Hence, as to the identification with the Tigris, 
the vocalism of the first syllable brings difficulties: it is i both in Old Persian and in 
Babylonian. Old Persian Tigram (with an alteration due to popular etymology, cf. 
Avestan tiyriS, Persian dr, "arrow") is borrowed from Babylonian Di-ik-lat (that 

Introduction 187 

rigid phonetic principles is the essential point, and means much more 
than any haphazardly made guesses at identification. Thus Mu-lu 
/fc }*&, name of a city on the eastern frontier of An-si (Parthia), 1 has 
been identified with Mouru (Muru, Merw) of the Avesta. 2 Whether 
this is historically correct, I do not wish to discuss here; from an his- 
torical viewpoint the identification may be correct, but from a phonetic 
viewpoint it is not acceptable, for Mu-lu corresponds to ancient *Muk- 
luk, Mug-ruk, Bug-luk, Bug-rug, to be restored perhaps to *Bux-rux. 8 
The scarcity of linguistic material on the Iranian side has imposed 
certain restrictions: names for Iranian plants, one of the chief subjects 
of this study, have been handed down to us to a very moderate extent, 
so that in many cases no identification can be attempted. I hope, 
however, that Iranian scholars will appreciate the philological con- 
tributions of the Chinese to Iranian and particularly Middle-Persian 
lexicography, for in almost every instance it is possible to restore with 
a very high degree of certainty the primeval Iranian forms from which 
the Chinese transcriptions were accurately made. The Chinese scholars 
had developed a rational method and a fixed system in reproducing 
words of foreign languages, in the study of which, as is well known, 
they took a profound interest; and from day to day, as our experience 
widens, we have occasion to admire the soundness, solidity, and con- 
sistency of this system. The same laws of transcription worked out 
for Sanskrit, Malayan, Turkish, Mongol, and Tibetan, hold good also 
for Iranian. I have only to ask Iranian scholars to have confidence in 
our method, which has successfully stood many tests. I am convinced 
that this plea is unnecessary for the savants of France, who are the 

is, Dik-lat, Dik-rat), which has passed into Greek Tlypijs and TLypis and Elamite 
Ti-ig-ra (A. Meillet, Grammaire du vieux perse, p. 72). It will thus be seen that 
the Chinese transcription *Dak-rat corresponds to Babylonian Dik-rat, save the 
vowel of the first element, which cannot yet be explained, but which will surely be 
traced some day to an Iranian dialect. — The T'ai p'in hwan yii ki (Ch. 185, p. 19) 
gives four geographical names of Persia, which have not yet been indicated. The 
first of these is the name of a city in the form ^ §| J§ Ho-p'o-kie, *Hat(r, 1)- 
bwa-g'iat. The first two elements *Har-bwa correspond to Old Persian Haraiva 
(Babylonian Hariva), Avestan Haraeva, Pahlavi *Harew, Armenian Hrew, — the 
modern Herat. The third element appears to contain a word with the meaning 
"city." The same character is used in $§j ^ J§'J Kie-li-pie, *G'iat-li-b'iet, name of a 
pass in the north-eastern part of Persia; here *g'iat, *g'iar, seems to represent 
Sogdian yr, *yara ("mountain"). Fan-tou ^ or ^ {ftj (Ts'ien Han $u, Ch. 96 a), 
anciently *Pan-tav, *Par-tav, corresponds exactly to Old Persian Parflava, Middle 
Persian Parflu. 

1 Hou Han Su, Ch. 116, p. 8 b. 

* Hirth, China and the Roman Orient, p. 143. 

• Cf. also the observation of E. H. Parker (Imp. and As. Quarterly Review, 
!903. p. 154)1 who noticed the phonetic difficulty in the proposed identification. 

1 88 Sino-Iranica 

most advanced and most competent representatives of the sinological 
field in all its varied and extensive branches, as well as in other domains 
of Oriental research. It would have been very tempting to summarize 
in a special chapter the Chinese method of transcribing Iranian and to 
discuss the phonology of Iranian in the light of Chinese contributions. 
Such an effort, however, appears to me premature at this moment: 
our knowledge of Sino-Iranian is in its infancy, and plenty of fresh 
evidence will come forward sooner or later from Turkistan manuscripts. 
There is no doubt that many hundreds of new Iranian terms of various 
dialects will be revived, and will considerably enrich our now scanty 
knowledge of the Iranian onomasticon and phonology. In view of the 
character of this publication, it was necessary to resort to a phonetic 
transcription of both ancient and modern Chinese on the same basis, 
as is now customary in all Oriental languages. The backwardness of 
Chinese research is illustrated by the fact that we slavishly adhere to 
a clumsy and antiquated system of romanization in which two and 
even three letters are wasted for the expression of a single sound. My 
system of transliteration will be easily grasped from the following com- 
parative table. 


ttg A 

Ch I 

CV r 

j & (while j serves to indicate the palatal 

I h $ sonant, written also dh). 

Other slight deviations from the old style, for instance, in the 
vowels, are self-explanatory. For the sake of the numerous compara- 
tive series including a large number of diverse Oriental languages it 
has been my aim to standardize the transcription as far as possible, 
with the exception of Sanskrit, for which the commonly adopted method 
remains. The letter x in Oriental words is never intended for the 
combination ks, but for the spirant surd, sometimes written kh. In 
proper names where we are generally accustomed to kh, I have allowed 
the latter to pass, perhaps also in other cases. I do not believe in super- 
consistency in purely technical matters. 

The linguistic phenomena, important as they may be, form merely 
a side-issue of this investigation. My main task is to trace the history 
of all objects of material culture, pre-eminently cultivated plants, 
drugs, products, minerals, metals, precious stones, and textiles, in their 
migration from Persia to China (Sino-Iranica), and others transmitted 
from China to Persia (Irano-Sinica). There are other groups of Sino- 
Iranica not included in this publication, particularly the animal world, 

Introduction 189 

games, and musical instruments. 1 The manuscript dealing with the 
fauna of Iran is ready, but will appear in another article the object of 
which is to treat all foreign animals known to the Chinese according 
to geographical areas and from the viewpoint of zoogeography in 
ancient and modern times. My notes on the games (particularly polo) 
and musical instruments of Persia adopted by the Chinese, as well as 
a study of Sino-Iranian geographical and tribal names, must likewise 
be reserved for another occasion. I hope that the chapter on the titles 
of the Sasanian government will be welcome, as those preserved in the 
Chinese Annals have been identified here for the first time. New 
results are also offered in the notice of Persian textiles. 

As to Iranian plants of which the Chinese have preserved notices, 
we must distinguish the following groups: (1) cultivated plants actually 
disseminated from Iranian to Chinese soil, (2) cultivated and wild 
plants of Iran merely noticed and described by Chinese authors, (3) drugs 
and aromatics of vegetable origin imported from Iran to China. The 
material, as far as possible, is arranged from this point of view and in 
chronological order. The single items are numbered. Apart from the 
five appendices, a hundred and thirty-five subjects are treated. At 
the outset it should be clearly understood that it is by no means the 
intention of these studies to convey the impression that the Chinese 
owe a portion of their material culture to Persia. Stress is laid on the 
point that the Chinese furnish us with immensely useful material for 
elaborating a history of cultivated plants. The foundation of Chinese 
civilization with its immense resources is no more affected by these 
introductions than that of Europe, which received numerous plants 
from the Orient and more recently from America. The Chinese merit 
our admiration for their far-sighted economic policy in making so 
many useful foreign plants tributary to themselves and amalgamating 
them with their sound system of agriculture. The Chinese were think- 
ing, sensible, and broad-minded people, and never declined to accept 
gratefully whatever good things foreigners had to offer. In plant- 
economy they are the foremost masters of the world, and China presents/ 
a unique spectacle in that all useful plants of the universe are cultivated 
there. Naturally, these cultivations were adopted and absorbed by a 
gradual process : it took the Chinese many centuries to become familiar 
with the flora of their own country, and the long series of their herbals 
(Pen ts'ao) shows us well how their knowledge of species increased 
from the T'ang to the present time, each of these works stating the 

1 Iranian influences on China in the matter of warfare, armor, and tactics have 
been discussed in Chinese Clay Figures, Part I. 

ioo Sino-Iranica 

number of additional species as compared with its predecessor. The 
introduction of foreign plants begins from the latter part of the second 
century B.C., and it was two plants of Iranian origin, the alfalfa and 
the grape-vine, which were the first exotic guests in the land of Han. 
These were followed by a long line of other Iranian and Central-Asiatic 
plants, and this great movement continued down to the fourteenth 
century in the Yuan period. The introduction of American species in 
the sixteenth and seventeenth centuries denotes the last phase in 
this economic development, which I hope to set forth in a special 
monograph. Aside from Iran, it was Indo-China, the Malayan region, 
and India which contributed a large quota to Chinese cultivations. 
It is essential to realize that the great Iranian plant-movement extends 
over a period of a millennium and a half; for a learned legend has been 
spread broadcast that most of these plants were acclimatized during 
the Han period, and even simultaneously by a single man, the well- 
known general, Can K'ien. It is one of my objects to destroy this 
myth. Can K'ien, as a matter of fact, brought to China solely two 
plants, — alfalfa and the grape-vine. No other plant is attributed to him 
in the contemporaneous annals. Only late and untrustworthy (chiefly 
Taoist) authors credit him also with the introduction of other Iranian 
plants. As time advanced, he was made the centre of legendary fabrica- 
tion, and almost any plant hailing from Central Asia and of doubtful 
or obscure history was passed off under his name: thus he was ulti- 
mately canonized as the great plant-introducer. Such types will 
spring up everywhere under similar conditions. A detailed discussion 
of this point will be found under the heading of each plant which by 
dint of mere fantasy or misunderstanding has been connected with 
Can K'ien by Chinese or European writers. In the case of the spinach 
I have furnished proof that this vegetable cannot have been culti- 
vated in Persia before the sixth century a.d., so that Can K'ien could 
not have had any knowledge of it. All the alleged Cah-K'ien plants 
were introduced into China from the third or fourth century a.d. down 
to the T'ang period inclusively (618-906). The erroneous reconstruction 
alluded to above was chiefly championed by Bretschneider and Hirth; 
and A. de Candolle, the father of the science of historical botany, who, 
as far as China is concerned, depended exclusively on Bretschneider, 
fell victim to the same error. 

F. v. Richthofen, 1 reproducing the long list of Bretschneider's 
Cah-K'ien plants, observes, "It cannot be assumed that Can K'ien 
himself brought along all these plants and seeds, for he had to travel 

1 China, Vol. I, p. 459. 

Introduction 191 

with caution, and for a year was kept prisoner by the Hiun-nu." When 
he adds, however, "but the relations which he had started brought the 
cultivated plants to China in the course of the next years," he goes on 
guessing or speculating. 

In his recent study of Can K'ien, Hirth 1 admits that of cultivated 
plants only the vine and alfalfa are mentioned in the Si ki. 2 He is 
unfortunate, however, in the attempt to safeguard his former position 
on this question when he continues to argue that "nevertheless, the one 
hero who must be looked upon as the pioneer of all that came from 
the West was Chang K'ien." This is at best a personal view, but an 
unhistorical and uncritical attitude. Nothing allows us to read more 
from our sources than they contain. The TsH min yao $u, to which 
Hirth takes refuge, can prove nothing -whatever in favor of his 
theory that the pomegranate, sesame, garlic, 3 and coriander were 
introduced by Can K'ien. The work in question was written at least 
half a millennium after his death, most probably in the sixth century 
a.d., and does not fall back on traditions coeval with the Han and 
now lost, but merely resorts to popular traditions evolved long after 
the Han period. In no authentic document of the Han is any allusion 
made to. any of these plants. Moreover, there is no dependence on 
the Ts'i min yao iw in the form in which we have this book at present. 
Bretschneider 4 said wisely and advisedly, "The original work was in 
ninety-two sections. A part of it was lost a long time ago, and much 
additional matter by later authors is found in the edition now cur- 
rent, which is in ten chapters. . . . According to an author of the 
twelfth century, quoted in the Wen hien fun k'ao, the edition then 
extant was already provided with the interpolated notes; and accord- 
ing to Li Tao, also an author of the Sung, these notes had been added 
by Sun Kuh of the Sung dynasty." 5 What such a work would be 
able to teach us on actual conditions of the Han era, I for my part 
am unable to see. 

1 Journal Am. Or. Soc, Vol. XXXVII, 1917, p. 92. The new translation of this 
chapter of the Si ki denotes a great advance, and is an admirable piece of work. It 
should be read by every one as an introduction to this volume. It is only on points 
of interpretation that in some cases I am compelled to dissent from Hirth's opinions. 

8 This seems to be the direct outcome of a conversation I had with the author 
during the Christmas week of 191 6, when I pointed out this fact to him and remarked 
that the alleged attributions to Can K'ien of other plants are merely the outcome of 
later traditions. 

* This is a double error (see below, p. 302). 

4 Bot. Sin., pt. I, p. 77. 

8 Cf. also Pelliot (Bull, de I'Ecole franqaise, Vol. IX, p. 434), who remarks, 
"Ce vieil et preaeux ouvrage nous est parvenu en assez mauvais 6tat." 

192 Sino-Iranica 

It has been my endeavor to correlate the Chinese data first of all 
with what we know from Iranian sources, and further with classical, 
Semitic, and Indian traditions. Unfortunately we have only fragments 
of Iranian literature. Chapter xxvn of the Bundahisn 1 contains a 
disquisition on plants, which is characteristic of the treatment of this 
subject in ancient Persia. As it is not only interesting from this point 
of view, but also contains a great deal of material to which reference 
will be made in the investigations to follow, an extract taken from 
E. W. West's translation 2 may be welcome. 

"These are as many genera of plants as exist: trees and shrubs, 
fruit-trees, corn, flowers, aromatic herbs, salads, spices, grass, wild 
plants, medicinal plants, gum plants, and all producing oil, dyes, and 
clothing. I will mention them also a second time: all whose fruit is 
not welcome as food of men, and are perennial, as the cypress, the 
plane, the white poplar, the box, and others of this genus, they call 
trees and shrubs (ddr va diraxt). The produce of everything welcome 
as food of men, that is perennial, as the date, the myrtle, the lote-plum 
(kiindr, a thorny tree, allied to the jujube, which bears a small plum- 
like fruit), the grape, the quince, the apple, the citron, the pomegranate, 
the peach, the fig, the walnut, the almond, and others in this genus, 
they call fruit (tnivak). Whatever requires labor with the spade, and 
is perennial, they call a shrub (diraxt). Whatever requires that they 
take its crop through labor, and its root withers away, such as wheat, 
barley, grain, various kinds of pulse, vetches, and others of this genus, 
they call corn (jurddk). Every plant with fragrant leaves, which is 
cultivated by the hand-labor of men, and is perennial, they call an 
aromatic herb (siparam). Whatever sweet-scented blossom arises at 
various seasons through the hand-labor of men, or has a perennial root 
and blossoms in its season with new shoots and sweet-scented blossoms, 
as the rose, the narcissus, the jasmine, the dog-rose (nestarun), the 
tulip, the colocynth (kavastik), the pandanus (kedi), the camba, the 
ox-eye (heri), the crocus, the swallow-wort (zarda), the violet, the 
karda, and others of this genus, they call a flower (gill). Everything 
whose sweet-scented fruit, or sweet-scented blossom, arises in its sea- 
son, without the hand-labor of men, they call a wild plant (vahdr or 
nihdl). Whatever is welcome as food of cattle and beasts of burden 
they call grass (giyah). Whatever enters into cakes (pes-pdrakihd) 
they call spices (dvzdrihd). Whatever is welcome in eating of bread, 
as torn shoots of the coriander, water-cress (kakij), the leek, and 

1 Cf. E. W. West, Pahlavi Literature, p. 98 (in Grundriss iran. Phil., Vol. II). 

2 Pahlavi Texts, pt. I, p. 100 (Sacred Books of the East, Vol V). 

Introduction 193 

others of this genus, they call salad (terak or tdrak, Persian tar ah). 
Whatever is like spinning cotton, and others of this genus, they call 
clothing plants (jdmak). Whatever lentil (mafag) is greasy, as sesame, 
dutedh, hemp, vandak (perhaps for zeto, 'olive,' as Anquetil supposes, 
and Justi assumes), and others of this genus, they call an oil-seed 
(rokano). Whatever one can dye clothing with, as saffron, sapan-wood, 
zalava, vaha, and others of this genus, they call a dye-plant (rag). 
Whatever root, or gum (tuf), or wood is scented, as frankincense 
(Pazand kendri for Pahlavi kundur), vardst (Persian barghast), kust, 
sandalwood, cardamom (Pazand kdkura, Persian qaqulah, 'cardamoms, 
or kakul, kdkul, 'marjoram')? camphor, orange-scented mint, and 
others of this genus, they call a scent (bod). Whatever stickiness 
comes out from plants they call gummy (vadak). The timber 
which proceeds from the trees, when it is either dry or wet, they 
call wood (Zihd). Every one of all these plants which is so, they call 
medicinal (ddruk). 

"The principal fruits are of thirty kinds, and there are ten species 
the inside and outside of which are fit to eat, as the fig, the apple, the 
quince, the citron, the grape, the mulberry, the pear, and others of this 
kind. There are ten the outside of which is fit to eat, but not the 
inside, as the date, the peach, the white apricot, and others of this kind; 
those the inside of which is fit to eat, but not the outside, are the walnut, 
the almond, the pomegranate, the coco-nut, 1 the filbert (Junduk), the 
chestnut (Sahbalut), the pistachio nut, the vargdn, and whatever else 
of this description are very remarkable. 

"This, too, it says, that every single flower is appropriate to an' 
angel (ametospend), 2 as the white jasmine (satnan) is for Vohuman, the 
myrtle and jasmine (ydsmin) are Auharmazd's own, the mouse-ear 
(or sweet marjoram) is A&avahist's own, the basil-royal is SatvirO's 
own, the musk flower is Spendarmad's, the lily is Horvadad's, the 
lamba is Amerodad's, Din-pavan-Ataro has the orange-scented mint 
(vddrang-bod), Ataro has the marigold (ddargun), the water-lily is 
A van's, the white marv is Xursed's, the ranges (probably rand, 'laurel') 
is Mah's, the violet is Tir's, the meren is Gos's, the kdrda is Din-pavan- 
Mitro's, all violets are Mitro's, the red chrysanthemum (xer) is Sr5§'s, 
the dog-rose (nestran) is Rasnu's, the cockscomb is Fravardin's, the 
sisebar is Vahram's, the yellow chrysanthemum is Ram's, the orange- 

1 Pazand anarsar is a misreading of Pahlavi anargll (Persian nargll), from 
Sanskrit ndrikela. 

2 These are the thirty archangels and angels whose names are applied to the 
thirty days of the Parsi month, in the order in which they are mentioned here, except 
that Auharmazd is the first day, and Vohuman is the second. 

194 Sino-Iranica 

scented mint is Vad's, the trigonella is Dln-pavan-Dln's, the hundred- 
petalled rose is Din's, all kinds of wild flowers (vahdr) are Ard's, Actad 
has all the white Horn, the bread-baker's basil is Asman's, Zamyad has 
the crocus, Maraspend has the flower of Ardaslr, Anlran has this 
Horn of the angel Horn, of three kinds." 

From this extract it becomes evident that the ancient Persians paid 
attention to their flora, and, being fond of systematizing, possessed a 
classification of their plants; but any of their botanical literature, if 
it ever existed, is lost. 

The most important of the Persian works on pharmacology is the 
Kitab-ulabniyat 'an haqd'iq-uladviyat or "Book of the Foundations of 
the True Properties of the Remedies," written about a.d. 970 by the 
physician Abu Mansur Muvaffaq bin 'All alharavi, who during one 
of his journeys visited also India. He wrote for Mansur Ibn Nuh II 
of the house of the Samanides, who reigned from 961 to 976 or 977. 
This is not only the earliest Persian work on the subject, but the 
oldest extant production in prose of New-Persian literature. The 
text has been edited by R. Seligmann from a unique manuscript 
of Vienna dated a.d. 1055, the oldest extant Persian manuscript. 1 
There is a translation by a Persian physician, Abdul-Chalig 
Achundow from Baku. 2 The translation in general seems good, and 
is provided with an elaborate commentary, but in view of the im- 
portance of the work a new critical edition would be desirable. 
The sources from which Abu Mansur derived his materials should 
be carefully sifted: we should like to know in detail what he 
owes to the Arabs, the Syrians, and the Indians, and what is due 
to his own observations. Altogether Arabic influence is pre-eminent. 
Cf. Appendix III. 

A good many Chinese plant-names introduced from Iran have the 
word Hu t$ prefixed to them. Hu is one of those general Chinese desig- 
nations without specific ethnic value for certain groups of foreign 
tribes. Under the Han it appears mainly to refer to Turkish tribes; 
thus the Hiun-nu are termed Hu in the Si ki. From the fourth century 
onward it relates to Central Asia and more particularly to peoples of 

1 Codex Vindobonensis sive Medici Abu Mansur Muwaffak Bin All Heratensis 
liber Fundamentorum Pharamacologiae Pars I Prolegomena et textum continens 
(Vienna, 1859). 

2 Die pharmakologischen Grundsatze des A. M. Muwaffak, in R. Kobert's 
Historische Studien aus dem Pharmakologischen Institute der Universitat Dorpat, 
l &73- Quoted as "Achundow, Abu Mansur." The author's name is properly 
'Abdu'l-Khaliq, son of the Akhund or schoolmaster. Cf. E. G. Browne, Literary 
History of Persia, pp. II, 478. 



Iranian extraction. 1 Bretschn eider 2 annotated, "If the character 
hu occurs in the name of a plant, it can be assumed that the plant is 
of foreign origin and especially from western Asia, for by Hu len the 
ancient Chinese denoted the peoples of western Asia." This is but 
partially correct. The attribute hu is by no means a safe criterion in 
stamping a plant as foreign, neither does hu in the names of plants 
which really are of foreign origin apply to West-Asiatic or Iranian 
plants exclusively. 

i. The word hu appears in a number of names of indigenous and 
partially wild plants without any apparent connection with the tribal 
designation Hu or without allusion to their provenience from the Hu. 
In the Li Sao, the famous elegies by K'u Yuan of the fourth century 
B.C., a plant is mentioned under the name hu Sen. #1 $81 , said to be a 
fragrant grass from which long cords were made. This plant is not 
identified. 3 

2. The acid variety of yu fib (Citrus grandis) is styled hu kan 
#J "HY apparently an ironical nickname, which may mean "sweet like 
the Hu." The tree itself is a native of China. 

3. The term hu hien ffl fL occurs only in the T'u kin pen ts'ao of 
Su Sun of the eleventh century as a variety of hien (Amarantus) , which 
is indigenous to China. It is not stated that this variety came from 
abroad, nor is it known what it really was. 

4. Hu mien man #J M ^ is a variety of Rehmannia, 5 a native 
of China and Japan. The name possibly means "the man with the face 
of a Hu." 6 C'en Ts'ah-k'i of the T'ang says in regard to this plant that 
it grows in Lih-nan (Kwah-tuh), and is like ti hwan #& iic (Rehmannia 

5. The pla^it known as ku-sui-pu H* #$ $f (Poly podium fortunei) 
is indigenous to China, and, according to C'en Ts'an-k'i, was called 

1 "Le terme est bien en principe, vers Tan 800, une designation des Iraniens et 
en particulier des Sogdiens " (Chavannes and Pelliot, Traite manichden, p. 231). 
This in general is certainly true, but we have well authenticated instances, traceable 
to the fourth century at least, of specifically Iranian plants the names of which are 
combined with the element Hu, that can but apply to Iranians. 

1 Chinese Recorder, 1 871, p. 22 1. 

8 Bretschneider, Bot. Sin., pt. II, No. 420; and Li sao Is'ao tnu su (Ch. 2, 
p. 16 b, ed. of Ci pu tsu £ai ts'un su) by Wu Zen-kie ^| £l §J! of the Sung period. 
See also T'ai pHn yii Ian, Ch. 994, p. 6 b. 

4 Bretschneider, op. cit., No. 236; W. T. Swingle in Plantas Wilsoniana?, 
Vol. 11, p. 130. 

* Stuart, Chinese Materia Medica, p. 372. 

6 Cf. analogous plant-names like our Jews-mallow, Jews-thorn, Jews-ear, Jews- 

196 Sino-Iranica 

by the people of Kian-si ffl J£ ^ hu-sun-kian, a purely local name 
which does not hint at any relation to the Hu. 

6. Another botanical name in which the word hu appears without 
reference to the Hu is Vui-hu-ken ^1 #3 J8, unidentified, a wild plant 
diffused all over China, and first mentioned by C'en Ts'an-k'i as grow- 
ing in the river-valleys of Kian-nan. 1 

7-8. The same remark holds good for ts'e-hu j£ (^) #l 2 (Bupleurum 
falcatum), a wild plant of all northern provinces and already described 
in the Pie lu, and for ts x ien-hu Hf fifl 3 {Angelica decursiva), growing in 
damp soil in central and northern China. 

9. Su-hu-lan 13 #} fl& is an unidentified plant, first and solely men- 
tioned by C'en Ts'ah-k'i, 4 the seeds of which, resembling those of 
Pimpinella anisum, are eatable and medicinally employed. It grows 
in Annam. One might be tempted to take the term as hu-lan of Su 
(Se-2'wan), but lu-hu-lan may be the transcription of a foreign word. 

10. The ma-k'in E$ jfjf or niu *r k'in (Viola pinnata), a wild violet, 
is termed hu k'in #J 7r in the T'un li 3® iS by Ceh Tsiao % Wl (1 108-62) 
and in the T'u kin pen ts'ao of Su Sufi. 6 No explanation as to the mean- 
ing of this hu is on record. 

11. The hu-man (wan) S3 s is a poisonous plant, identified with 
Gelsemium elegans.* It is mentioned in the Pei hu lu 1 with the synonyme 
ye-ko Jn H, 8 the vegetable yun ^g (Ipomoea aquatica) being regarded as 
an antidote for poisoning by hu-man. C'en Ts'ah-k'i is cited as au- 
thority for this statement. The Lin piao lu i 9 writes the name I? 35, 
and defines it as a poisonous grass; hu-man grass is the common col- 
loquial name. The same work further says, "When one has eaten of 
this plant by mistake, one should use a broth made from sheep's blood 
which will neutralize the poison. According to some, this plant grows 
as a creeper. Its leaves are like those of the Ian hian S3 #, bright and 
thick. Its poison largely penetrates into the leaves, and is not employed 

1 Pen ts'ao kan tnu, Ch. 16, p. 7 b. 
• *Op. cit., Ch. 13, p. 6 b. 

* Op. cit., Ch. 13, p. 7 b. 

* Op. cit., Ch. 26, p. 22 b. 

* Op. cit., Ch. 26, p. 21 ; Ci wu min Si t'u k'ao, Ch. 14, p. 76. 

•Cf. C. Ford, China Review, Vol. XV, 1887, pp. 215-220. Stuart (Chinese 
Materia Medica, p. 220) says that the plant is unidentified, nevertheless he describes 
it on p. 185. 

7 Ch. 2, p. 18 b (ed. of Lu Sin-yoan). 

8 According to Matsumura (Shokubutsu mei-i, No. 2689), Rhus toxicodendron 
(Japanese tsuta-uruii). 

9 Ch. B, p. 2 (ed. of Wu yin Hen). 

Introduction 197 

as a drug. Even if an antidote is taken, this poison will cause death 
within a half day. The goats feeding on the sprouts of this plant will 
fatten and grow." Fan C'en-ta t& $c ^C (1126-93), i n his Kwei hai 
yii hen Zi, x mentions this plant under the name hu-man t'en j§£ ("hu-man 
creeper"), saying that it is a poisonous herb, which, rubbed and soaked 
in water, will result in instantaneous death as soon as this liquid enters 
the mouth. The plant is indigenous to southern China, and no reason 
is given for the word hu being prefixed to it. 

12. Hu Vui-tse ~$ M ■? (literally, "chin of the Hu") is the name 
of an evergreen tree or shrub indigenous throughout China, even to 
Annam. The name is not explained, and there are no data in Chinese 
records to indicate that it was introduced from abroad. 2 It is men- 
tioned by C'en Ts'ah-k'i as a tree growing in P'in-lin *P $s and it is 
said to be alluded to in the chapter Wu hin ci 3l fx i& of the Sun s"u. 
The synonyme kHo'r-su ^ jS 8£ ("sparrow-curd," because the birds 
are fond of the fruit) first appears in the Pao ci lun of Lei Hiao of the 
fifth century. The people of Yue call the plant p'u-Vui-tse W M "& ; 
the southerners, lu-tu-tse M. HE ■?, which according to Liu Tsi ^J ^ 
of the Ming, in his Fei sue lu IS W- $k, is a word from the speech of 
the Man. The people of Wu term the tree pan-han-c'un ^ff$, 
because its fruit ripens at an early date. The people of Sian IS style 
it hwah-p'o-nai JI9$flR ("yellow woman's breast"), because the 
fruit resembles a nipple. 

13. In hu-lu 5§8 or HI iH. (Lagenaria vulgaris) the first character is 
a substitute for 1$ hu. The gourd is a native of China. 

14. Hui-hui tou IhJ j£ (literally, "Mohammedan bean") is a 
plant everywhere growing wild in the fields. 8 The same remark holds 
good for hu tou W[ i£, a kind of bean which is roasted or made into 
flour, according to the Pen ts'ao H i, a weed growing in rice-fields. Wu 
K'i-ts'un, author of the Ci wu min H Vu k'ao, says, "What is now hu tou, 
grows wild, and is not the hu tou of ancient times." 4 

15. Yen hu su $j£ #i ^ denotes tubers of Corydalis ambigua: they 
are little, hard, brown tubers, of somewhat flattened spherical form, 
averaging half an inch in diameter. The plant is a native of Siberia, 

1 Ed. of Ci pu tsu lai ts'un Su, p. 30. 

1 Stuart (Chinese Materia Medica, p. 161) is mistaken in saying that several 
names of this plant are "possibly transliterations of Turkic or Mongol names." 
There are no such names on record. The tree is identified with Elcsagnus longipes 
or pungens. 

3 Ci wu min Si t'u k'ao, Ch. 2, p. 11 b. _It is first mentioned in the Kiu hwan 
Pen ts'ao, being also called na-ho-tou |)5 ^ B. 

4 See, further, below, p. 305. 

198 Sino-Iranica 

Kamchatka, and the Amur region, and flowers upon the melting of the 
snow in early spring. 1 According to the Pen ts'ao kan mu, 2 the plant 
is first mentioned by C'en Ts'an-k'i of the T'ang period as growing in 
the country Hi f&, and came from Nan-tun $t M (in Korea). Li Si-cen 
annotates that by Hi the north-eastern barbarians should be under- 
stood. Wan Hao-ku 3: jft ii, a physician of the thirteenth century, 
remarks that the name of the plant was originally hiian 5£ hu-su, but 
that on account of a taboo (to avoid the name of the Emperor Cen-tsun 
of the Sung) it was altered into yen-hu-su; but this explanation cannot 
be correct, as the latter designation is already ascribed to C'en Ts'an-k'i 
of the T'ang. It is not known whether hu in this case would allude to 
the provenience of the plant from Korea. In the following example, 
however, the allusion to Korea is clear. 

The mint, W ffi po-ho, *bak-xa (Mentha arvensis or aquatica), occurs 
in China both spontaneously and in the cultivated state. The plant 
is regarded as indigenous by the Chinese, but also a foreign variety is 
known as hu pa-ho (*bwat-xa) #! ^ jlfi. 3 C'en §i-liaft ffl> ~il It, in his 
Si sin pen ts'ao 'ffc'144^, published in the tenth century, introduced 
the term wu i$| pa-ho, "mint of Wu" (that is, Su-cou, where the best 
mint was cultivated), in distinction from hu pa-ho, "mint of the Hu." 
Su Sun, in his T'u kin pen ts'ao, written at the end of the eleventh 
century, affirms that this foreign mint is similar to the native species, 
the only difference being that it is somewhat sweeter in taste; it grows 
on the border of Kian-su and Ce-kian, where the people make it 
into tea; commonly it is styled Sin-lo M H po-ho, "mint of Sinra" 
(in Korea). Thus this variety may have been introduced under the 
Sung from Korea, and it is to this country that the term hu may refer. 

Li §i-2en relates that Sun Se-miao 3% M> 28, in his Ts'ien kin fan 
*r & jfr, 4 writes the word ^ W fan-ho, but that this is erroneously due 
to a dialectic pronunciation. This means, in other words, that the first 
character fan is merely a variant of ^, 5 and, like the latter, had the 
phonetic equivalent *bwat, bat. 6 

1 Hanbury, Science Papers, p. 256. 

2 Ch. 13, p. 13. 

' The word po-ho is Chinese, not foreign. The Persian word for "peppermint" 
is pudene, pudina, budenk (Kurd punk) ; in Hindi it is pudlna or pudinekd, derived 
from the Persian. In Tibetan (Ladakh) it is p'o-lo-lin; in the Tibetan written lan- 
guage, byi-rug-pa, hence Mongol jirukba; in Manchu it is farsa. 

* See below, p. 306. 

6 As Sun Se-miao lived in the seventh century, when the Korean mint was not 
yet introduced, his term fan-ho could, of course, not be construed to mean "foreign 

• In T'oung Pao (1915, p. 18) Pelliot has endeavored to show that the char- 

Introduction 199 

In the following example there is no positive evidence as to the 
significance of hu. Hu wan Si Ze ffl 3E ^ 41 (" envoy of the king of the 
Hu") is a synonyme of tu hwo ffl ?§ (Peucedanum decursivum). 1 As 
the same plant is also styled k'ian tsHh 7^ W , k'ian hwo, and hu k'ian 
H ce Ml 7& $i %, the term K'ian (*Gian) alluding to Tibetan tribes, it 
may be inferred that the king of the Hu likewise hints at Tibetans. 
In general, however, the term Hu does not include Tibetans, and the 
present case is not conclusive in showing that it does. In the chapter 
on the walnut it will be seen that there are two introduced varieties, — 
an Iranian (hu t'ao) and a Tibetan one (k'ian t'ao). 

In hu ts'ai (Brassica rapa) the element hu, according to Chinese 
tradition, relates to Mongolia, while it is very likely that the vegetable 
itself was merely introduced there from Iran. 2 

In other instances, plants have some relation to the Hu; but what 
this relation is, or what group of tribes should be understood by Hu, 
is not revealed. 

There is a plant, termed hu hwan lien SH !fc 31, the hwan-lien (Coptis 
teeta) of the Hu, because, as Li Si-cen says, its physical characteristics, 
taste, virtue, and employment are similar to those of hwan-lien. It 
has been identified with Barkhausia repens. As evidenced by the 

acter fan, on the authority of K'ah-hi, could never have had the pronunciation po 
nor a final consonant, and that, accordingly, in the tribal name T'u-fan (Tibet) the 
character fan, as had previously been assumed, could not transcribe the Tibetan 
word bod. True it is that under the character in question K'an-hi has nothing to 
say about po, but ^§p is merely a graphic variant of f£, with which it is phonetically 
identical. Now under this character, K'ah-hi indicates plainly that, according to the 
Tsi yiin and Cen yun, fan in geographical names is to be read p'o (anciently *bwa) 
§j£ (fan-ls'ie $f $£), and that, according to the dictionary Si wen, the same char- 
acter was pronounced P'o (*bwa) §£, p'u |jff , and p'an§R£(ci. also Schlegel, Secret of 
the Chinese Method, pp. 21-22). In the ancient transcription K| or £ 9fi fan-tou, 
*par-tav, reproduction of Old Persian Parflava (see above, p. 1 87) ,fan corresponds very 
well to par or bar; and if it could interchange with the phonetic f£ pa, *bwat, bwar, it is 
perfectly clear that, contrary to Pelliot's theory, there were at least dialectic cases, 
where ^ was possessed of a final consonant, being sounded bwat or bwar. Con- 
sequently it could have very well served for the reproduction of Tibetan bod. From 
another phonetic viewpoint the above case is of interest: we have *bak-xa and 
*bwat-xa as ancient names for the mint, which goes to show that the final con- 
sonants of the first element were vacillating or varied in different dialects (cf . Voung 
Pao, 1916, pp. 110-114). 

1 "Pun ci (above, p. 196), Ch. 75, p. 12 b. 

2 See below, p. 381. In the term hu yen ("swallow of the Hu")» Am appears to 
refer to Mongolia, as shown by the Manchu translation monggo cibin and the TurkI 
equivalent qalmaq qarlogal (Mongol xatun xariyatsai, Tibetan gyi-gyi k'ug-rta; cf. 
Ross, Polyglot List of Birds, No. 267). The bird occurs not only in Mongolia, but 
also in Ce-kian Province, China (see Kwei ki sanfu iu "f^ f§ E£ |K jet, Ch. 2, p. 8; 
ed. of Si yin hiian ts'un §u). 

200 Sino-Iranica 

attribute Hu, it may be of foreign origin, its foreign name being H>J ffl> 
% W ko-hu-lu-tse (*kat-wu-lou-dzak). Unfortunately it is not indicated 
at what time this transcription was adopted, nor does Li Si-cen state 
the source from which he derived it. The only T'ang author who 
mentions the plant, Su Kun, does not give this foreign name. At all 
events, it does not convey the impression of representing a T'ang 
transcription; on the contrary, it bears the ear-marks of a transcription 
made under the Yuan. Su Kun observes, "Hu hwan-lien is produced 
in the country Po-se and grows on dry land near the sea-shore. Its 
sprouts are like those of the hia-ku ts'ao Jt $T ^ (Brunella vulgaris). 
The root resembles a bird's bill; and the cross-section, the eyes of the 
mainah. The best is gathered in the first decade of the eighth month." 
Su Sun of the Sung period remarks that the plant now occurs in Nan-hai 
(Kwah-tun), as well as in Ts'in-lun §HPti (Sen-si and Kan-su). This 
seems to be all the information on record. 1 It is not known to me that 
Barkhausia grows in Persia; at least, Schlimmer, in his extensive dic- 
tionary of Persian plants, does not note it. 

$ou-ti M. ■(£ is mentioned by C'en Ts'an-k'i as a plant (not yet 
identified) with seeds of sweet and warm flavor and not poisonous, and 
growing in Si-fan (Western Barbarians or Tibet) and in northern China 
At zh, resembling hwai Man ^ ^r {Pimpinella anisum). The Hu make 
the seeds into a soup and eat them. 2 In this case the term Hu may be 
equated with Si-fan, but among the Chinese naturalists the latter term 
is somewhat loosely used, and does not necessarily designate Tibet. 3 

Hiun-k'iun *§f || (Conioselinum univittatum) is an umbelliferous 
plant, which is a native of China. As early as the third century a.d. 
it is stated in the Wu H pen ts'ao 4 that some varieties of this plant grow 
among the Hu; and Li Si-cen annotates that the varieties from the Hu 
and Zun are excellent, and are hence styled hu k'iuh tft |f. 5 It is stated 
that this genus is found in mountain districts in Central Europe, 
Siberia, and north-western America. 6 N 

1 What Stuart (Chinese Materia Medica, p. 65) says regarding this plant is 
very inexact. He arbitrarily identifies the term Hu with the Kukunor, and wrongly 
ascribes Su Kuh's statement to T'ao Huh-kih. Such an assertion as, "the drug is 
now said to be produced in Nan-hai, and also in Sen-si and Kan-su," is misleading, 
as this "now" comes from an author of the Sung period, and does not necessarily 
hold good for the present time. 

2 Pen ts'ao kan mu, Ch. 26, p. 22 b. 

3 Cf . below, p. 344. 

4 Cf . Beginnings of Porcelain, p. 115. 

5 He also imparts a Sanskrit name from the Suvarnaprabhasa-sutra in the form 
f§ Jt£ 7& Se-mo-k'ie, *ja-mak-gia. The genus is not contained in Watt's Dictionary. 

6 Treasury of Botany, Vol. I, p. 322. 

Introduction * 201 


In hu tsiao ("pepper") the attribute hu distinctly refers to India. 1 
Another example in which hu alludes to India is presented by the 
term hu kan kian U\ $L ^ ("dried ginger of the Hu"), which is a 
synonyme of T'ien-tu % *5 kan kian ("dried ginger of India"), "pro- 
duced in the country of the Brahmans." 2 

In the term hufen tfi $r (a cosmetic or facial powder of white lead), 
the element hu bears no relation to the Hu, although it is mentioned 
as a product of Kuca 3 and subsequently as one of the city of Ili (Yi-li- 
pa-li). 4 In fact, there is no Chinese tradition to the effect that this 
substance ever came from the Hu. 6 F. P. Smith 8 observed with refer- 
ence to this subject, "The word hu does not denote that the substance 
was formerly obtained from some foreign source, but is the result of a 
mistaken character." This evidently refers to the definition of the 
dictionary $i min W & by Liu Hi of the Han, who explains this hu 
by t$ hu ("gruel, congee"), which is mixed with grease to be rubbed 
into the face. The process of making this powder from lead is a thor- 
oughly Chinese affair. 

In the term hu yen ffl IS ("salt of the Hu") the word Hu refers to 
barbarous, chiefly Tibetan, tribes bordering on China in the west; for 
there are also the synonymes Hun -& yen and k x ian 3& yen, the former 
already occurring in the Pie lu. Su Kun of the seventh century equalizes 
the terms Hun yen and hu yen, and gives t'u-ten 35 $£ yen as the word 
used in Sa-cbu & /H. Ta Mih ^C *5§, who wrote in a.d. 970, says that this 
is the salt consumed by the Tibetans (Si-fan), and hence receives the 
designation Hun or k'ian yen. Other texts, however, seem to make a 
distinction between hu yen and Hun yen: thus it is said in the biography 
of Li Hiao-po ^ # f 9 in the Wei Su, "The salt of the Hu cures pain 
of the eye, the salt of the 2un heals ulcers." 

The preceding examples are sufficient to illustrate the fact that 
the element hu in botanical terms demands caution, and that each case 
must be judged on its own merits. No hard and fast rule, as deduced 
by Bretschneider, can be laid down: the mere addition of hu proves 
neither that a plant is foreign, nor that it is West-Asiatic or Iranian. 
There are native plants equipped with this attribute, and there are 
foreign plants thus characterized, which hail from Korea, India, or 

1 See below, p. 374. 

2 Cen lei pen ts'ao, Ch. 6, p. 67 b. 

* Cou $u, Ch. 50, p. 5; Sui Su, Ch. 83, p. 5 b. 

* Ta Min i t'un li, Ch. 89, p. 22; Kwan yii ki, Ch. 24, p. 6 b. 

' Pen ts'ao kan mu, Ch. 8, p. 6; Geerts (Produits, pp. 596-601), whose transla- 
tion "poudre des pays barbares" is out of place. 

* Contributions towards the Materia Medica of China, p. 231. 

202 Sino-Iranica 

some vaguely defined region of Central Asia. The fact, however, re- 
mains that there are a number of introduced, cultivated Hu plants 
coming from Iranian lands, but in each and every case it has been my 
endeavor to furnish proof for the fact that these actually represent 
Iranian cultivations. With the sole exception of the walnut, the his- 
tory of which may tolerably well be traced, the records of these Hu 
plants are rather vague, and for none of them is there any specific 
account of the introduction. It is for botanical rather than historical 
reasons that the fact of the introduction becomes evident. It is this 
hazy character of the traditions which renders it impossible to connect 
these plants in any way with Can K'ien. Moreover, it cannot be 
proved with certainty that any names of plants or products formed 
with the element hu existed under the Han. The sole exception would 
be hu ts'ai, 1 but its occurrence in the T'un su wen of the Han is not 
certain either; and this hu, according to Chinese tradition, refers to 
Mongolia, not to Iran. Another merely seeming exception is presented 
by hu t'un-lei* but this is a wild, not a cultivated tree; and hu, in this 
case, has a geographical rather than an ethnographical significance. In 
the wooden documents discovered in Turkistan we have one good, 
datable instance of a Hu product; and this is hu Vie ("iron of the Hu" 
and implements made of such iron). These tablets belong to the Tsin 
period (a.d. 265-419),* while in no wooden document of the Han has 
any compound with Hu as yet been traced. Again, all available evi- 
dence goes to show that these Hu plants were not introduced earlier 
than the Tsin dynasty, or, generally speaking, during what is known 
as the Leu c'ao or six minor dynasties, covering the time from the 
downfall of the Han to the rise of the T'ang dynasty. It is noteworthy 
that of none of these plants is an Iranian name on record. 

The element hu, in a few cases, serves also the purpose of a tran- 
scription: thus probably in the name of the coriander, hu-swi,* and 
quite evidently in the name of the fenugreek, hu-lu-pa} 

Imported fruits and products have been named by many nations 
for the countries from which they hailed or from the people by whom 
they were first brought. The Greeks had their "Persian apple" (jxrj\op 
Ucpaindv, "peach"), their "Medic apple" (wXov Mt;5ik6v, "citron"), 
their "Medic grass" (MtjSiKi) 71-60, "alfalfa"), and their "Armenian 

1 Below, p. 381. I 

s Below, p. 339. 

* Chavannes, Documents chinois ddcouverts par Aurel Stein, pp. 168, 169. 
4 Below, p. 298. 

1 Below, p. 446. It thus occurs also in geographical names, as in Hu-2'a-la 
(Guzerat); see Hirth and Rockhill, Chao Ju-kua, p. 92. 

Introduction 203 

apple" (hjj\ov 'Apfxeviandv, "apricot"). Rabelais (1483-1553) 1 has 
already made the following just observation on this point, " Les autres 
[plantes] ont retenu le nom des regions des quelles furent ailleurs 
transporters, comme pommes medices, ce sont pommes de Medie, en 
laquelle furent premierement trouv^es; pommes puniques, ce sont 
grenades, apportees de Punicie, c'est Carthage. Ligusticum, c'est 
livesche, apport£e de Ligurie, c'est la couste de Genes: rhabarbe, du 
fleuve Barbare nomine" Rha, comme atteste Ammianus: santonique, 
fenu grec; castanes, persiques, sabine; stoechas, de mes isles Hieres, 
antiquement dites Stoechades; spica celtica et autres." The Tibetans, 
as I have shown, 2 form many names of plants and products with Bal 
(Nepal), Mon (Himalayan Region), rGya (China), and Li (Khotan). 

In the same manner we have numerous botanical terms preceded 
by "American, Indian, Turkish, Turkey, Guinea," etc. 

Aside from the general term Hu, the Chinese characterize Iranian 
plants also by the attribute Po-se (Parsa, Persia): thus Po-se tsao 
("Persian jujube") serves for the designation of the date. The term 
Po-se requires great caution, as it denotes two different countries, Persia 
and a certain Malayan region. This duplicity of the name caused 
grave confusion among both Chinese and European scholars, so that 
I was compelled to devote to this problem a special chapter in which 
all available sources relative to the Malayan Po-se and its products 
are discussed. Another tribal name that quite frequently occurs in 
connection with Iranian plant-names is Si-2un f§ & ("the Western 
£un"). These tribes appear as early as the epoch of the Si kin and 
$u kin, and seem to be people of Hiun-nu descent. In post-Christian 
times Si-zun developed into a generic term without ethnic significance, 
and vaguely hints at Central-Asiatic regions. Combined with botanical 
names, it appears to be synonymous with Hu. 3 It is a matter of course 
that all these geographical and tribal allusions in plant-names have 
merely a relative, not an absolute value; that is, if the Chinese, for 
instance, designate a plant as Persian (Po-se) or Hu, this signifies that 
from their viewpoint the plant under notice hailed from Iran, or in 
some way was associated with the activity of Iranian nations, but it 
does not mean that the plant itself or its cultivation is peculiar or due 
to Iranians. This may be the case or not, yet this point remains to be 
determined by a special investigation in each particular instance. 
While the Chinese, as will be seen, are better informed on the history 

1 Le Gargantua et le Pantagruel, Livre III, chap. L. 

* "Poung Poo, 1916, pp. 409, 448, 456. , 

* For examples of its occurrence consult Index. 

204 Sino-Iranica 

of important plants than any other people of Asia (and I should even 
venture to add, of Europe), the exact and critical history of a plant- 
cultivation can be written only by heeding all data and consulting all 
sources that can be gathered from every quarter. The evidence accruing 
from the Semites, from Egypt, Greece, and Rome, from the Arabs, 
India, Camboja, Annam, Malayans, Japan, etc., must be equally 
requisitioned. Only by such co-ordination may an authentic result be 
hoped for. 

The reader desirous of information on the scientific literature 
of the Chinese utilized in this publication may be referred to Bret- 
schneider's "Botanicon Sinicum" (part I). 1 It is regrettable that no 
Pen ts'ao (Herbal) of the T'ang period has as yet come to light, and 
that for these works we have to depend on the extracts given in later 
books. The loss of the Hu pen ts'ao ("Materia Medica of the Hu") 
and the C'u hu kwo fan ("Prescriptions from the Hu Countries") is 
especially deplorable. I have directly consulted the Cen lei pen ts'ao, 
written by T'ah Sen-wei in 1108 (editions printed in 1521 and 1587), 
the Pen ts'ao yen i by K'ou Tsun-si of 11 16 in the edition of Lu Sin- 
yuan, and the well-known and inexhaustible Pen ts'ao kan mu by Li 
Si-cen, completed in 1578. With all its errors and inexact quotations, 
this remains a monumental work of great erudition and much solid 
information. Of Japanese Pen ts'ao (Honzo) I have used the Yamato 
honzo, written by Kaibara Ekken in 1709, and the Honzo komoku keimo 
by Ono Ranzan. Wherever possible, I have resorted to the original 
source-books. Of botanical works, the Kwah k'iinfan p'u, the Hwa p'u, 
the Ci wu min Si t'u k'ao, and several Japanese works, have been utilized. 
The Yu yah tsa tsu has yielded a good many contributions to the plants 
of Po-se and Fu-lin; several Fu-lin botanical names hitherto unexplained 
I have been able to identify with their Aramaic equivalents. Although 
these do not fall within the subject of Sino-Iranica, but Sino-Semitica, 
it is justifiable to treat them in this connection, as the Fu-lin names 
are given side by side with the Po-se names. Needless to say, I have 
carefully read all accounts of Persia and the Iranian nations of Central 
Asia contained in the Chinese Annals, and the material to be found 
there constitutes the basis and backbone of this investigation. 2 

There is a class of literature which has not yet been enlisted for the 

1 We are in need, however, of a far more complete and critical history of the 
scientific literature of the Chinese. 

2 The non-sinological reader may consult to advantage E. H. Parker, Chinese 
Knowledge of Early Persia (Imp. and Asiatic Quarterly Review, Vol. XV, 1903, 
pp. 144-169) for the general, contents of the documents relating to Persia. Most 
names of plants and other products have been omitted in Parker's article. 

Introduction 205 

study of cultivated plants, and this is the early literature on medicine. 
Prominent are the books of the physician Can Cun-kin 5H ity Hk or 
Can Ki 35 JS, who is supposed to have lived under the Later Han at 
the end of the second century a.d. A goodly number of cultivated plants 
is mentioned in his book Kin kwei yii han yao lio fan lun 4£ fSf 3E #3 
3c H ~J5 Ift or abbreviated Kin kwei yao lio. 1 This is a very interesting 
hand-book of dietetics giving detailed rules as to the avoidance of 
certain foods at certain times or in certain combinations, poisonous 
effects of articles of diet, and prescriptions to counteract this poison. 
Neither this nor any other medical writer gives descriptions of plants 
or notes regarding their introduction; they are simply enumerated in 
the text of the prescriptions. But it is readily seen that, if such a work 
can be exactly dated, it has a chronological value in determining whether 
a given plant was known at that period. Thus Can Eli mentions, of 
plants that interest us in this investigation, the walnut, the pome- 
granate, the coriander, and Allium scorodoprasum (hu swan). Unfortu- 
nately, however, we do not know that we possess his work in its 
original shape, and Chinese scholars admit that it has suffered from inter- 
polations which it is no longer possible to unravel. The data of such 
a work must be utilized with care whenever points of chronology are 
emphasized. It was rather tempting to add to the original prescrip- 
tions of Can Ki, and there is no doubt that the subsequent editions 
have blended primeval text with later comments. The earliest com- 
mentary is by Wan Su-ho 3: $£ ^P of the Tsin. Now, if we note that 
the plants in question are otherwise not mentioned under the Han, but 
in other books are recorded only several centuries later, we can hardly 
refrain from entertaining serious doubts as to Can Ki's acquaintance 
with them. A critical bibliographical study of early Chinese medical 
literature is an earnest desideratum. 

A. de Candolle's monumental work on the "Origin of Cultivated 
Plants" is still the only comprehensive book on this subject that we 
have. It was a masterpiece for his time, and still merits being made 
the basis and starting-point for any investigation of this kind. De Can- 
dolle possessed a really critical and historical spirit, which cannot be 
said of other botanists who tried to follow him on the path of his- 
torical research; and the history of many cultivated plants has been 
outlined by him perfectly well and exactly. Of many others, our con- 
ceptions are now somewhat different. Above all, it must be said that 

1 Reprinted in the Yu tswan i Isun kin kien of 1739 (Wylie, Notes on Chinese 
Literature, p. 101). A good edition of this and the other works of the same author on 
the basis of a Sung edition is contained in the medical Ts'ufi-su, the J t'un ten mo 
ts'uan Su, published by the Ce-kian Su ku. 

206 Sino-Iranica 

since his days Oriental studies have made such rapid strides, that his 
notes with regard to India, China, and Japan, are thoroughly out of 
date. As to China, he possessed no other information than the super- 
ficial remarks of Bretschneider in his "Study and Value of Chinese 
Botanical Works," 1 which teem with misunderstandings and errors. 2 
De Candolle's conclusions as to things Chinese are no longer acceptable. 
The same holds good for India and probably also for Egypt and western 
Asia. In point of method, de Candolle has set a dangerous precedent 
to botanists in whose writings this effect is still visible, and this is 
his over-valuation of purely linguistic data. The existence of a native 
name for a plant is apt to prove little or nothing for the history of 
the plant, which must be based on documentary and botanical evi- 
dence. Names, as is well known, in many cases are misleading or 
deceptive; they constitute a welcome accessory in the chain of evidence, 
but they cannot be relied upon exclusively. It is a different case, of 
course, if the Chinese offer us plant-names which can be proved to be 
of Iranian origin. If on several occasions I feel obliged to uphold 
V. Hehn against his botanical critic A. Engler, such pleas must not 
be construed to mean that I am an unconditional admirer of Hehn; 
on the contrary, I am wide awake to his weak points and the short- 
comings of his method, but wherever in my estimation he is right, it 
is my duty to say that he is right. A book to which I owe much in- 
formation is Charles Joret's "Les Plantes dans l'antiquite' et au 
moyen age" (2 vols., Paris, 1897, 1904), which contains a sober and 
clear account of the plants of ancient Iran. 3 

A work to which I am greatly indebted is " Terminologie medico- 
pharmaceutique et anthropologique francaise-persane, " by J. L. 
Schlimmer, lithographed at Teheran, 1874. 4 This comprehensive work 
of over 600 pages folio embodies the lifelong labors of an instructor at 
the Polytechnic College of Persia, and treats in alphabetical order of 
animal and vegetable products, drugs, minerals, mineral waters, native 

1 Published in the Chinese Recorder for 1870 and 1871. 

1 They represent the fruit of a first hasty and superficial reading of the Pen 
ts'ao kan mu without the application of any criticism. In Chinese literature we can 
reach a conclusion only by consulting and sifting all documents bearing on a problem. 
Bretschneider's Botanicon Sinicum, much quoted by sinologues and looked upon as 
a sort of gospel by those who are unable to control his data, has now a merely relative 
value, and is uncritical and unsatisfactory both from a botanical and a sinological 
viewpoint; it is simply a translation of the botanical section of the Pen ts'ao kan mu 
without criticism and with many errors, the most interesting plants being omitted. 

1 Joret died in Paris on December 26, 1914, at the age of eighty-five years 
(cf. obituary notice by H. Cordier, La Geographie, 19 14, p. 239). 

* Quoted "Schlimmer, Terminologie." I wish to express my obligation to the 
Surgeon General's Library in Washington for the loan of this now very rare book. 

Introduction 207 

therapeutics and diseases, with a wealth of solid information that has 
hardly ever been utilized by our science. 

It is hoped that these researches will chiefly appeal to botanists 
and to students of human civilization; but, as it can hardly be expected 
that the individual botanist will be equally interested in the history 
of every plant here presented, each subject is treated as a unit and 
as an independent essay, so that any one, according to his inclination 
and choice, may approach any chapter he desires. Repetitions have 
therefore not been shunned, and cross-references are liberally inter- 
spersed; it should be borne in mind, however, that my object is not 
to outline merely the history of this or that plant, but what I wish to 
present is a synthetic and comprehensive picture of a great and unique 
plant-migration in the sense of a cultural movement, and simultane- 
ously an attempt to determine the Iranian stratum in the structure of 
Chinese civilization. It is not easy to combine botanical, oriental, 
philological, and historical knowledge, but no pains have been spared 
to render justice to both the botanical and the historical side of each 
problem. All data have been sifted critically, whether they come 
from Chinese, Japanese, Indian, Persian, Arabic, or classical sources, 
and in no instance have I depended on a second-hand or dogmatic 
statement. The various criticisms of A. de Candolle, A. Engler, E. 
Bretschneider, and other eminent authorities, arise from the critical 
attitude toward the subject, and merely aim at the furtherance of the 

I wish to express my thanks to Dr. Tanaka TyOzaburO in the 
Bureau of Plant Industry of the Department of Agriculture, Washing- 
ton, for having kindly prepared a translation of the notices on the 
grape-vine and the walnut from Japanese sources, which are appended 
to the chapters on the history of these plants. The manuscript of this 
publication was completed in April, 191 8. 

The generosity of Mrs. T. B. Blackstone and Mr. Charles R. 
Crane in contributing a fund toward the printing of this volume is 
gratefully acknowledged. 


I. The earliest extant literary allusion to alfalfa 1 (Medicago sativa) 
is made in 424 B.C. in the Equites ("The Knights") of Aristophanes, 
who says (V, 606) : 

"Hadiov 81 roiis irayobpovs o\vtI irolas nrjSiKrjs. 

"The horses ate the crabs of Corinth as a substitute for the Medic.*! 

The term "Medike" is derived from the name of the country Media. 
In his description of Media, Strabo* states that the plant constituting 
the chief food of the horses is called by the Greeks "Medike" from its 
growing in Media in great abundance. He also mentions as a product 
of Media silphion, from which is obtained the Medic juice. 3 Pliny 4 
intimates that "Medica" is by nature foreign to Greece, and that it 
was first introduced there from Media in consequence of the Persian 
wars under King Darius. Dioscorides 6 describes the plant without 
referring to a locality, and adds that it is used as forage by the cattle- 
breeders. In Italy, the plant was disseminated from the middle of the 
second century B.C. to the middle of the first century a.d., 8 — almost 
coeval with its propagation to China. The Assyriologists claim that 
aspasti or aspastu, the Iranian designation of alfalfa, is mentioned in 
a Babylonian text of ca. 700 B.C.; 7 and it would not be impossible that 
its favorite fodder followed the horse at the time of its introduction 
from Iran into Mesopotamia. A. de Candolle 8 states that Medicago 

1 1 use this term (not lucerne) in accordance with the practice of the U. S. 
Department of Agriculture; it is also the term generally used and understood by the 
people of the United States. The word is of Arabic origin, and was adopted by the 
Spaniards, who introduced it with the plant into Mexico and South America in the 
sixteenth century. In 1854 it was taken to San Francisco from Chile (J. M. West- 
gate, Alfalfa, p. 5, Washington, 1908). \ 

1 XI. xni, 7. 

• Theophrastus (Hist, plant., VIII. vn, 7) mentions alfalfa but casually by 
saying that it is destroyed by the dung and urine of sheep. Regarding silphion 
see p. 355. 

4 xni, 43. 

•11, 176. 

6 Hehn, Kulturpflanzen, 8th ed., p. 412. 

T Schrader in Hehn, p. 416; C. Joret (Plantes dans l'antiquite\ Vol. II, p. 68) 
states after J. Hal6vy that aspasti figures in the list drawn up by the gardener of the 
Babylonian king Mardukbalidin (Merodach-Baladan), a contemporary of Ezechias 
Bang of Juda. 

8 Origin of Cultivated Plants, p. 103. 


Alfalfa 209 

sativa has been found wild, with every appearance of an indigenous 
plant, in several provinces of Anatolia, to the south of the Caucasus, 
in several parts of Persia, in Afghanistan, Baluchistan, and in Kashmir. 1 
Hence the Greeks, he concludes, may have introduced the plant from 
Asia Minor as well as from India, which extended from the north of 
Persia. This theory seems to me inadmissible and superfluous, for 
the Greeks allude solely to Media in this connection, not to India. 
Moreover, the cultivation of the plant is not ancient in India, but is 
of recent date, and hardly plays any r61e in Indian agriculture and 

In ancient Iran, alfalfa was a highly important crop closely associated 
with the breeding of superior races of horses. Pahlavi aspast or aspist 
New Persian aspust, uspust, aspist, ispist, or isfist (Pu§tu or Afghan spastu, 
SpeZta), is traceable to an Avestan or Old-Iranian *aspo-asti (from the 
root ad, "to eat"), and literally means "horse-fodder." 2 This word has 
penetrated into Syriac in the form aspesta or pespestd (the latter in the 
Geoponica). Khosrau I (a.d. 531-578) of the Sasanian dynasty included 
alfalfa in his new organization of the land-tax: 3 the tax laid on alfalfa 
was seven times as high as that on wheat and barley, which gives an 
idea of the high valuation of that forage-plant. It was also employed 
in the pharmacopoeia, being dealt with by Abu Mansur in his book 
on pharmacology. 4 The seeds are still used medicinally. 5 The Arabs 
derived from the Persians the word isfist, Arabicized into fisfisa; Arabic 
designations being ratba and qatt, the former for the plant in its natural 
state, the latter for the dried plant. 6 

The mere fact that the Greeks received Medicago from the Persians, 
and christened it "Medic grass," by no means signifies or proves at the 
outset that Medicago represents a genuinely Iranian cultivation. It is 
well known how fallacious such names are: the Greeks also had the 
peach under the name "Persian apple," and the apricot as "Armenian 
apple;" yet peach and apricot are not originally Persian or Armenian, 
but Chinese cultivations: Iranians and Armenians in this case merely 

1 As to Kashmir, it will be seen, we receive a confirmation from an ancient 
Chinese document. See also G. Watt, Dictionary of the Economic Products of 
India, Vol. V, pp. 199-203. 

2 Neldeke, ZDMG, Vol. XXXII, 1878, p. 408. Regarding some analogous 
plant-names, see R. v. Stackelberg, ibid., Vol. LIV, 1900, pp. 108, 109. 

8 Noldeke, Tabari, p. 244. 

4 Achundow, Abu Mansur, p. 73 (cf. above, p. 194). 

6 Schlimmer, Terminologie, p. 365. He gives yond£e as the Persian name, which, 
however, is of Turkish origin (from yont, "horse"). In Asia Minor there is a place 
Yonjali ("rich in alfalfa"). 

6 Leclerc, Traite" des simples, Vol. Ill, p. 35. 

210 Sino-Iranica 

acted as mediators between the far east and the Mediterranean. How- 
ever, the case of alfalfa presents a different problem. The Chinese, who 
cultivate alfalfa to a great extent, do not claim it as an element of 
their agriculture, but have a circumstantial tradition as to when and 
how it was received by them from Iranian quarters in the second 
century B.C. As any antiquity for this plant is lacking in India or any 
other Asiatic country, the verdict as to the centre of its primeval culti- 
vation is decidedly in favor of Iran. The contribution which the Chinese 
have to make to the history of Medicago is of fundamental importance 
and sheds new light on the whole subject: in fact, the history of no 
cultivated plant is so well authenticated and so solidly founded. 

In the inscription of Persepolis, King Darius says, "This land Persia 
which Auramazda has bestowed on me, being beautiful, populous, and 
abundant in horses — according to the will of Auramazda and my own, 
King Darius — it does not tremble before any enemy." I have alluded 
in the introduction to the results of General Can K'ien's memorable 
expedition to Central Asia. The desire to possess the fine Iranian 
thoroughbreds, more massively built than the small Mongolian horse, 
and distinguished by their noble proportions and slenderness of feet 
as well as by the development of chest, neck, and croup, was one of 
the strongest motives for the Emperor Wu (140-87 B.C.) to maintain 
regular missions to Iranian countries, which led to a regular caravan 
trade with Fergana and Parthia. Even more than ten such missions 
were dispatched in the course of a year, the minimum being five or six. 
At first, this superior breed of horse was obtained from the Wu-sun, 
but then it was found by Can K'ien that the breed of Fergana was far 
superior. These horses were called "blood-sweating" (han-hue ff jfiL), 1 
and were believed to be the offspring of a heavenly horse {Vien ma 
% fi$). The favorite fodder of this noble breed consisted in Medicago 
sativa; and it was a sound conclusion of General Can K'ien, who was a 
practical man and possessed of good judgment in economic matters, 
that, if these much-coveted horses were to continue to thrive on Chinese 
soil, their staple food had to go along with them. Thus he obtained 
the seeds of alfalfa in Fergana, 2 and presented them in 126 B.C. to his 
imperial master, who had wide tracts of land near his palaces covered 

1 This name doubtless represents the echo of some Iranian mythical concept, 
but I have not yet succeeded in tracing it in Iranian mythology. 

2 In Fergana as well as in the remainder of Russian Turkistan Medicago sativa 
is still propagated on an immense scale, and represents the only forage-plant of that 
country, without which any economy would be impossible, for pasture-land and hay 
are lacking. Alfalfa yields four or five harvests there a year, and is used for the feed- 
ing of cattle either in the fresh or dry state. In the mountains it is cultivated up to 
an elevation of five thousand feet; wild or as an escape from cultivation it reaches 

Alfalfa 211 

with this novel plant, and enjoyed the possession of large numbers of 
celestial horses. 1 From the palaces this fodder-plant soon spread to 
the people, and was rapidly diffused throughout northern China. 
According to Yen Si-ku (a.d. 570-645), this was already an accom- 
plished fact during the Han period. As an officinal plant, alfalfa appears 
in the early work Pie lu. 1 The TsH min yao iu of the sixth century 
a.d. gives rules for its cultivation; and T'ao Hun-kin (a.d. 451-536) 
remarks that "it is grown in gardens at C'an-han (the ancient capital 
in Sen-si), and is much valued by the northerners, while the people 
of Kiah-nan do not indulge in it much, as it is devoid of flavor. Abroad 
there is another mu-su plant for healing eye-diseases, but different 
from this species." 3 

Can K'ien was sent out by the Emperor Wu to search for the 
Yue-£i and to close an alliance with them against the Turkish Hiun-nu. 
The Yue-6i, in my opinion, were an Indo-European people, speaking a 
North-Iranian language related to Scythian, Sogdian, YagnObi, and 
Ossetic. In the course of his mission, Can K'ien visited Fergana, Sog- 
diana, and Bactria, all strongholds of an Iranian population. The 
"West" for the first time revealed by him to his astounded country- 
men was Iranian civilization, and the products which he brought back 
were thoroughly and typically Iranian. The two cultivated plants 
(and only these two) introduced by him into his fatherland hailed 
from Fergana: Ferganian was an Iranian language; and the words for 
the alfalfa and grape, mu-su and p'u-Vao, were noted by Can K'ien 
in Fergana and transmitted to China along with the new cultivations. 
These words were Ferganian; that is, Iranian. 4 Can K'ien himself was 

an altitude up to nine thousand feet. Cf. S. Korzinski, Vegetation of Turkistan 
(in Russian), p. 51. Russian Turkistan produces the largest supply of alfalfa-seed 
for export (E. Brown, Bull. Dep. of Agriculture, No. 138, 1914). 

1 Si ki, Ch. 123. 

1 Cf. Chinese Clay Figures, p. 135. 

8 Cen lei pen ts'ao, Ch. 27, p. 23. It is not known what this foreign species is. 

1 Hirth's theory (Journal Am. Or. Soc, Vol. XXXVII, 1917, p. 149), that the 
element yuan of Ta-yuan (Fergana) might represent a "fair linguistic equivalent" of 
Yavan (Yavana, the Indian name of the Greeks), had already been advanced by J. 
Edkins {Journal China Branch Roy. As. Soc, Vol. XVIII, 1884, p. 5). To me it 
6eems eccentric, and I regret being unable to accept it. In the T'ang period we have 
from Huan Tsan a reproduction of the name Yavana in the form $Q |$f $5 
Yen-mo-na, *Yam-mwa-na (Pelliot, Bull, de I'Ecole francaise, Vol. IV, p. 278). 
For the Han period we should expect, after the analogy of ^ M Ye-tiao, *Yap 
(Dzap)-div (Yavadvlpa, Java), a transcription |j| JJ5 Ye-na, *Yap-na, for Yavana. 
The term jfe j|j§ Yu-yue, *Yu-vat(var), does not represent a transcription of Yavana, 
as supposed by Chavannes (Memoires historiques de Se-ma Ts'ien, Vol. IV, 1901, 
pp. 558-559), but is intended to transcribe the name Yuan (*Yuvar, Yu,ar), 
Btill employed by the Cam and other peoples of Indo-China as a designation of 

212 Sino-Iranica 

very well aware of the fact that the speech of the people of Fergana was 
Iranian, for he stated in his report, that, although there were different 
dialects in the tract of land stretching from Fergana westward as far 
as Parthia (An-si), yet their resemblance was so great that the people 
could make themselves intelligible to each other. 1 This is a plain 
allusion to the differentiation and at the same time the unity of Iranian 
speech; 2 and if the Ferganians were able to understand the Parthians, 
I do not see in what other language than Iranian they could have 
conversed. Certainly they did not speak Greek or Turkish, as some 
prejudiced theorists are inclined to imagine. 

The word brought back by Can K'ien for the designation of alfalfa, 
and still used everywhere in China for this plant, was mu-su H fit, 
consisting of two plain phonetic elements, 8 anciently *muk-suk (Japa- 
nese moku-Suku), subsequently written H" Hf with the addition of the 
classifier No. 140. I recently had occasion to indicate an ancient Tibetan 
transcription of the Chinese word in the form bug-sug, 4 and this appears 
to come very near to the Iranian prototype to be restored, which was 
*buksuk or *buxsux, perhaps *buxsuk. The only sensible explanation 
ever given of this word, which unfortunately escaped the sinologues, 
was advanced by W. Tomaschek, 6 who tentatively compared it with 
Gllakl (a Caspian dialect) buso ("alfalfa"). This would be satisfactory 
if it could be demonstrated that this buso is evolved from *bux-sox or 
the like. Further progress in our knowledge of Iranian dialectology 

Annam and the Annamese (cf. Cam Yuan or Yuon, Bahnar, Juon, Khmer Yuon, 
Stien Ju6n). This native name, however, was adapted to or assimilated with Sanskrit 
Yavana; for in the Sanskrit inscriptions of Campa, particularly in one of the reign 
of Jaya-Rudravarman dated a.d. 1092, Annam is styled Yavana (A. Bergaigne, 
L'Ancien royaume de Campa, p. 61 of the reprint from Journal asiatique, 1888). 
In the Old-Javanese poem NagarakrtSgama, completed in a.d. 1365, Yavana 
occurs twice as a name for Annam (H. Kern ,Bijdragen totde taal- land- envolkenkunde, 
Vol.LXXII, 1916, p. 399). Kern says that the question as to how the name of the 
Greeks was applied to Annam has not been raised or answered by any one; he over- 
looked the contribution of Bergaigne, who discussed the problem. 

1 Strabo (XV. n, 8) observes, "The name of Ariana is extended so as to include 
some part of Persia, Media, and the north of Bactria and Sogdiana; for these peoples 
speak nearly the same language." 

2 Emphasized by R. Gauthiot in his posthumous work Trois Memoires sur 
l'unit6 linguistique des parlers iraniens (reprinted from the Memoires de la Societe" 
de Linguistique de Paris, Vol. XX, 1916). 

3 The two characters are thus indeed written without the classifiers in the Han 
Annals. The writings tfc j^f *muk-suk of Kwo P'o and 7^ ^| *muk-swok of Lo 
Yuan, author of the Er ya i (simply inspired by attempts at reading certain mean- 
ings into the characters), have the same phonetic value. In Annamese it is muk-tuk. 

* T'oung Pao, 1916, p. 500, No. 206. 

6 Pamir-Dialekte (Sitzber. Wiener Akad., 1880, p. 792). 

Alfalfa 213 

will no doubt supply the correct form of this word. We have to be 
mindful of the fact that the speech of those East-Iranian tribes, the 
advance-guard of Iran proper, with whom the Chinese first came in 
contact, has never been committed to writing, and is practically lost 
to us. Only secluded dialects may still harbor remnants of that lost 
treasure. We have to be the more grateful to the Chinese for having 
rescued for us a few words of that extinct language, and to place *buksuk 
or *buxsux on record as the ancient Ferganian appellation of Medicago 
sativa. The first element of this word may survive in Sariqoll (a Pamir 
dialect) wux ("grass"). In WaxI, another Pamir idiom, alfalfa is 
styled wujerk; and grass, wiiL "Horse" is ya$ in WaxI, and vurj in 
Sariqoll. 1 

Bretschneider 2 was content to say that mu-su is not Chinese, 
but most probably a foreign name. Watters, in his treatment of 
foreign words in Chinese, has dodged this term. T. W. Kingsmill 8 
is responsible for the hypothesis that mu-su "may have some connec- 
tion with the M7761K17 PoTavrj of Strabo." This is adopted by the Chinese 
Dictionary of Giles.* This Greek designation had certainly not pene- 
trated to Fergana, nor did the Iranian Ferganians use a Greek name 
for a plant indigenous to their country. It is also impossible to see 
what the phonetic coincidence between *muk-suk or *buk-suk and 
medike is supposed to be. 

The least acceptable explanation of mu-su is that recently pro- 
pounded by Hirth, 6 who identifies it with a Turkish burfak, which is 
Osmanli, and refers to the pea. 8 Now, it is universally known that a 
language like Osmanli was not in existence in the second century B.C., 
but is a comparatively modern form of Turkish speech; and how Can 
K'ien should have picked up an Osmanli or any other Turkish word for 
a typically Iranian plant in Fergana, where there were no Turks at that 
time, is unintelligible. Nor is the alleged identification phonetically 
correct: Chinese mu, *muk, *buk, cannot represent bur, nor can su, 

1 Cf. R. B. Shaw, On the Ghalchah Languages (Journal As. Soc. Bengal, 1876, 
pp. 221, 231). According to Tomaschek (op. cit., p. 763), this word is evolved from 
*bharaka, Ossetic bairag ("good foal"). 

2 Bot. Sin., pt. Ill, p. 404. 

8 Journal China Branch Roy. As. Soc, Vol. XIV, 1879, p. 19. 

4 No. 8081, wrongly printed MeSiK-rj. The word Pot&vt) is not connected with 
the name of the plant, but in the text of Strabo is separated from Mij5ik^»> by eleven 
words. Mij5i/c^ is to be explained as scil. w6a, "Medic grass or fodder." 

6 Journal Am. Or. Soc, Vol. XXXVII, 1917, p. 145. 

6 Kara burlak means the "black pea" and denotes the vetch. 

214 Sino-Iranica 

*suk, stand for Zak} The entire speculation is deplorable, and we are 
even expected "to allow for a change the word may have undergone 
from the original meaning within the last two thousand years"; but 
there is no trace of evidence that the Osmanli word has existed that 
length of time, neither can it be reasonably admitted that the signifi- 
cance of a word can change from "pea" to "alfalfa." The universal 
term in Central Asia for alfalfa is bidd 2 or beda, z Djagatai bidd. This 
word means simply "fodder, clover, hay." 4 According to Tomaschek, 6 
this word is of Iranian origin (Persian beda). It is found also in Sariqoll, 
a Pamir dialect. 6 This would indicate very well that the Persians 
(and it could hardly be expected otherwise) disseminated the alfalfa 
to Turkistan. 

According to Vambery, 7 alfalfa appears to have been indigenous 
among the Turks from all times; this opinion, however, is only based 
on linguistic evidence, which is not convincing: a genuine Turkish 
name exists in Djagatai jonuSka (read yonucka) and Osmanli yondza* 
(add Kasak-Kirgiz yonurcka), which simply means "green fodder, 
clover." Now, these dialects represent such recent forms of Turkish 
speech, that so far-reaching a conclusion cannot be based on them. 
As far as I know, in the older Turkish languages no word for alfalfa 
has as yet been found. 

A Sanskrit §§ M- fJ jM sai-pi-li-k'ie , *sak-bi-lik-kya, for the designa- 
tion of mu-su, is indicated by Li Si-cen, 9 who states that this is the 
word for mu-su used in the Kin kwan min kin -§£%!$% W± (Suvar- 
naprabhasa-sutra). This is somewhat surprising, in view of the fact 
that there is no Sanskrit word for this plant known to us; 10 and there 
can be no doubt that the latter was introduced into India from Iran 
in comparatively recent times. Bretschneider's suggestion, 11 that in 

1 Final k in transcriptions never answers to a final r, but only to k, g, or x (cf. 
also Pelliot, T'oung Pao, 1912, p. 476). V 

2 A. Stein, Khotan, Vol. I, p. 130. 

8 Le Coq, SprichwSrter und Lieder aus Turfan, p. 85. 

4 I. Kunos, Sulejman Efendi's Cagataj-Osman. Worterbuch, p. 26. 

6 Pamir- Dialekte, p. 792. 

8 R. B. Shaw, Journal As. Soc. Bengal, 1876, p. 231. 

7 Primitive Cultur des turko-tatarischen Volkes, p. 220. 

8 The etymology given of this word by VambSry is fantastic and unacceptable. 

9 Pen ts'ao kan mu, Ch. 27, p. 3 b. Mu-su is classified by him under ts'ai 

10 This was already remarked by A. de Candolle (Origin of Cultivated Plants, 
p. 104). Also Watt gives only modern Indian vernacular names, three of which, 
spastu, sebist, and beda, are of Iranian origin. 

u Bot. Sin., pt. Ill, p. 404. 

Alfalfa 215 

Kabul the Trifolium giganteum is called sibarga, and Medicago sativa 
is styled rilka, is unsatisfactory. The word sibarga means "trefoil" 
(si, "three;" barga = Persian barak, varak, "leaf"), and is Iranian, not 
Sanskrit; the corresponding Sanskrit word is tripatra or triparna. The 
word rilka is Afghan; that is, likewise Iranian. 1 Considering the fact 
that nothing is known about the plant in question in early Indian 
sources, it is highly improbable that it should figure in a Buddhist 
Sutra of the type of the Suvarnaprabhasa; and I think that Li Si-5en 
is mistaken as to the meaning of the word, which he says he encountered 

The above transcription occurs also in the Fan yi min yi tsi 
(section 27) and answers to Sanskrit gaka-vrika, the word caka denoting 
any eatable herb or vegetable, and vfika (or baka) referring to a certain 
plant not yet identified (cf. the analogous formation qaka-bilva, "egg- 
plant"). It is not known what herb is to be understood by qdka-vfika, 
and the Chinese translation mu-su may be merely a makeshift, though 
it is not impossible that the Sanskrit compound refers to some species 
of Medicago. We must not lose sight of the fact that the equations 
established in the Chinese-Sanskrit dictionaries are for the greater part 
merely bookish or lexicographical, and do not relate to plant introduc- 
tions. The Buddhist translators were merely anxious to find a suitable 
equivalent for an Indian term. This process is radically different from 
the plant-names introduced together with the plants from Iranian, 
Indian, or Southeast-Asiatic regions: here we face living realities, 
there we have to do with literary productions. Two other examples 
may suffice. The Fan yi min yi tsi (section 24) offers a Sanskrit botani- 
cal name in the form fl| W. j&I len-Vou-kia, anciently *tsin(tin)-du-k / ie, 
answering to Sanskrit tinduka (Diospyros embryopteris) , a dense ever- 
green small tree common throughout India and Burma. The Chinese 
gloss explains the Indian word by Si ffi, which is the well-known Dio- 
spyros kaki of China and Japan, not, however, found in ancient India; it 
was but recently introduced into the Botanical Garden of Calcutta by 
Col. Kyd, and the Chinese gardeners employed there call it lin ("Chi- 
nese"). 2 In this case it signifies only the Diospyros embryopteris of 
India. Under the heading kan-sun hian (see p. 455), which denotes the 
spikenard (Nardostachys jatamansi), Li Si-Sen gives a Sanskrit term 
i^r ffl B£ k'u-mi-Fe, *ku-mi-£i, likewise taken from the Suvarnapra- 
bhasasatra; this corresponds to Sanskrit kunci or kuncika, which applies 
to three different plants, — 1. Abrus precatorius, 2. Nigella indica, 

1 There are, further, in Afghan sebist (connected with Persian supust) and 

8 W. Roxburgh, Flora Indica, p. 412. 

216 Sino-Iranica 

3. Trigonella f&num graecum. In this case the compromise is a failure, 
or the identification of kunci with kan-sun even results from an error; 
the Sanskrit term for the spikenard is gandhamamsi. 

We must not draw inferences from mere Sanskrit names, either, as to 
the origin of Chinese plants, unless there is more substantial evidence. 
Thus Stuart 1 remarks under li ^ {Prunus domestica) that the Sanskrit 
equivalent M ^ $E kii-lih-kia indicates that this plum may have been 
introduced from India or Persia. Prunus domestica, however, is a native 
of China, mentioned in the Si kin, Li ki, and in Mon-tse. The Sino- 
Indian word is given in the Fan yi min yi tsi (section 24) with the trans- 
lation li. The only corresponding Sanskrit word is kulinga, which 
denotes a kind of gall. The question is merely of explaining a Sanskrit 
term to the Chinese, but this has no botanical or historical value for the 
Chinese species. 

Thus the records of the Chinese felicitously supplement the meagre 
notices of alfalfa on the part of the ancients, and lend its history 
the proper perspective: we recognize the why and how of the world- 
wide propagation of this useful economic plant.' Aside from Fergana, 
the Chinese of the Han period discovered mu-su also in Ki-pin (Kash- 
mir), 8 and this fact is of some importance in regard to the early geo- 
graphical distribution of the species; for in Kashmir, as well as in 
Afghanistan and Baluchistan, it is probably spontaneous. 4 

Mu-su gardens are mentioned under the Emperor Wu (a.d. 265-290) 
of the Tsin dynasty, and the post-horses of the T'ang dynasty were fed 
with alfalfa. 6 

The fact that alfalfa was used as an article of human food under 
the T'ang we note from the story of Sie Lin-<H I? & £., preceptor at 
the Court of the Emperor Yuan Tsun (a.d. 713-755), who wrote a 
versified complaint of the too meagre food allotted to him, in which 
alfalfas with long stems were the chief ingredient. 6 The good teacher, 
of course, was not familiar with the highly nutritive food-values of 
the plant. 

1 Chinese Materia Medica, p. 358. 

1 It is singular that A. de Candolle, in his Origin of Cultivated Plants, while he 
has conscientiously reproduced from Bretschneider all his plants wrongly ascribed 
to Can K'ien, does not make any reference to China in speaking of Medicago 
(pp. 102-104). In fact, its history has never before been outlined correctly. 

3 TsHen Han Su, Ch. 96 A. 

* A. de Candolle, op. cit., p. 103 ; G. T. Vigne, Travels in Kashmir, Vol. II, p. 455. 

6 S. Matsuda & EB /£ lK< 0° Medicago sativa and the Species of Medicago 
in China {Botanical Magazine fl| 4$? ^ $j| fji, Tokyo, Vol. XXI, 1907, p. 243). 
This is a very interesting and valuable study written in Japanese. 

a Cf . C. Petillon, Allusions litteraires, p. 350. 

Alfalfa 217 

According to the $u i ki *& H Ifi, written by Zen Fan & B# in 
the beginning of the sixth century, "the mu-su (alfalfa) gardens of 
Can K'ien are situated in what is now Lo-yan; mu-su was originally 
a vegetable in the land of the Hu, and K'ien was the first to obtain it 
in the Western Countries." A work, Kiu Vi ki ffc %& ffi, 1 says that east 
of the capital there were mu-su gardens, in which there were three 
pestles driven by water-power. 

The Si kin tsa ki 15 M H I2 2 states, "In the Lo-yu gardens M%tM 
(in the capital C'ah-nan) there are rose-bushes Jfc $&$$ (Rosa rugosa), 
which grow spontaneously. At the foot of these, there is abundance 
of mu-su, called also hwaifuh Hi US. ('embracing the wind'), sometimes 
kwah fun Jfc $&< ('brilliant wind'). 3 The people of Mou-lin ]3c Wt* style 
the plant lien-U ts'ao 3* t£ ^ ('herb with connected branches')." 6 

The Lo yah k'ie Ian ki $& 8» f&H III 12, a record of the Buddhist 
monasteries in the capital Lo-yan, written by Yan Huan-ci tfk $r 1<£. in 
a.d. 547 or shortly afterwards, says that "Huan-wu Jl 3£ is situated 
north-east of the Ta-hia Gate ^cSPI; now it is called Kwan-fun 
Garden Jt %, ^, producing mu-su." Kwan-fun, as shown by the Si kin 
tsa ki, is a synonyme of mu-su. 

K'ou Tsuh-Si, in his Pen ts'ao yen i, a written in a.d. 1116, notes that 
alfalfa is abundant in Sen-si, being used for feeding cattle and horses, 
and is also consumed by the population, but it should not be eaten in 
large quantity. Under the Mongols, the cultivation of alfalfa was 
much encouraged, especially in order to avert the danger of famines; 7 
and gardens were maintained to raise alfalfa for the feeding of horses. 8 
According to Li Si-cen (latter part of the sixteenth century), 9 it was in 
his time a common, wild plant in the fields everywhere, but was culti- 
vated in Sen-si and Kan-su. He apparently means, however, Medicago 
denticulata, which is a wild species and a native of China. Forbes 

1 T'ai p'ift yii Ian, Ch. 824, p. 9. 

2 That is, Miscellaneous Records of the Western Capital (C'an-nan in Sen-si), 
written by Wu Kiin jj| $£] of the sixth century a.d. 

1 The explanation given for these names is thus: the wind constantly whistles 
in these gardens, and the sunlight lends brilliancy to the flowers. 

4 Ancient name for the present district of Hin-p'in ^ ^ in the prefecture of 
Si-nan, Sen-si. 

B T'ai p'in yii Ian, Ch. 996, p. 4 b. 

6 Ch. 19, p. 3 (ed. of Lu Sin-yuan). 

7 Yuan Si, Ch. 93, p. 5 b. 

8 Ibid., Ch. 91, p. 6 b. 

9 Pen ts'ao kan mu, Ch. 28, p. 3 b. 

«i8 Sino-Iranica 

and Hemsley 1 give as Chinese species Medicago denticulata, falcata, 1 
and lupulina (the black Medick or nonsuch), M. lupulina "apparently 
common, and from the most distant parts," and say with reference to 
Medicago sativa that it is cultivated in northern China, and also occurs 
in a wild state, though it is probably not indigenous. This "wild" 
Medicago saliva may be an escape from cultivation. It is an interesting 
point that those wild species are named ye mu-su ("wild alfalfa"), 
which goes to show that these were observed by the Chinese only after 
the introduction of the imported cultivated species. 3 Wu K'i-tsun 4 
has figured two ye mu-su, following his illustration of the mu-su, — one 
being Medicago lupulina, the other M. denticulata. 

The Japanese call the plant uma-goyali ("horse-nourishing"). 6 
Matsumura 6 enumerates four species: M. sativa: murasaki ("purple") 
umagoyaU; 1 M. denticulata: umagoyaH; M. lupulina: kometsubu- 
umagoyaH; and M. minima: ko-umagoyaH. 

In the Tibetan dialect of Ladakh, alfalfa is known as ol. This word 
refers to the Medicago sativa indigenous to Kashmir or possibly intro- 
duced there from Iran. In Tibet proper the plant is unknown. In 
Armenia occur Medicago sativa, M. falcata, M. agrestis, and M. 
lupulina. 8 

Under the title "Notice sur la plante mou-sou ou luzerne chinoise 
par C. de Skattschkoff, suivie d'une autre notice sur la menie plante 
traduite du chinois par G. Pauthier," a brief article of 16 pages appeared 
in Paris, 1864, as a reprint from the Revue de I 'Orient. 9 Skattschkoff, 
who had spent seven years in Peking, subsequently became Russian 
consul in Dsungaria, and he communicates valuable information on the 
agriculture of Medicago in that region. He states that seeds of this 

1 Journal Linnean Soc, Vol. XXIII, p. 154. 

1 Attempts are being made to introduce and to cultivate this species in the 
United States (cf. Oakley and Garver, Medicago Falcata, U. S. Department of 
Agriculture, Bull. No. 428, 1917). 

8 We shall renew this experience in the case of the grape-vine and the walnut. 

4 Ci wu min H t'u k'ao, Ch. 3, pp. 58, 59. 

8 In the same manner, Manchu morxo is formed from morin ("horse") and 
orxo ("grass"). 

• Shoku butsu-mei-i, Nos. 183-184. 

T The flower of this species is purple-colored. 

8 A. Beguinot and P. N. Diratzsuyan, Contributo alia flora dell' Armenia, 

9 The work of Pauthier is limited to a translation of the notice on the plant in 
the Ci wu min Si t'u k'ao. The name Yu-lou nun frequently occurring in this work 
does not refer to a treatise on agriculture, as conceived by Pauthier, but is the literary 
style of Wu K'i-tsun, author of that work. 

Alfalfa 219 

plant were for the first time sent from China to Russia in 1840, and 
that he himself has been active for six years in propagating it in Russia, 
Livonia, Esthonia, and Finland. This is not to be doubted, but the 
point I venture to question is that the plant should not have been 
known in Russia prior to 1840. Not only do we find in the Russian 
language the words medunka (from Greek medike) and the European 
Vutserna (lucerne) for the designation of Medicago sativa, but also 
krasni ("red") burkun, lefuxa, lugovoi v'azel ("Coronilla of the 
meadows"); the word burkun, burundHk, referring to Medicago falcata 
(called also y&morki), burunlik to M. lupulina. It is hard to realize 
that all these terms should have sprung up since 1840, and that the 
Russians should not have received information about this useful plant 
from European, Iranian, or Turkish peoples. A. de Candolle 1 ob- 
serves, "In the south of Russia, a locality mentioned by some authors, 
it is perhaps the result of cultivation as well as in the south of Europe." 
Judging from the report of N. E. Hansen, 1 it appears that three species 
of Medicago (M. falcata, M. platycarpa, and M. ruthenica) are indigenous 
to Siberia. 

The efforts of our Department of Agriculture to promote and to 
improve the cultivation of alfalfa in this country are well known; for 
this purpose also seeds from China have been introduced. Argentine 
chiefly owes to alfalfa a great amount of its cattle-breeding. 8 

1 Origin of Cultivated Plants, p. 103. 

8 The Wild Alfalfas and Clovers of Siberia, pp. 1 1-15 (Bureau of Plant Industry, 
Bull. No. 150, Washington, 1909). 

* Cf. I. B. Lorenzetti, La Alfafa en la Argentina (Buenos Aires, 1913, 360 p.)» 


2. The grape-vine (Vitis vinifera) belongs to the ancient cultivated 
plants of western Asia and Egypt. It is not one of the most ancient 
cultivations, for cereals and many kinds of pulse are surely far earlier, 
but it is old enough to have its beginnings lost in the dawn of history. 
Viticulture represents such a complexity of ideas, of a uniform and 
persistent character throughout the ancient world, that it can have 
been disseminated but from a single centre. Opinions as to the loca- 
tion of this focus are of course divided, and our present knowledge of 
the subject does not permit us to go beyond more or less probable 
theories. Certain it is that the primeval home of vine-growing is to 
be sought in the Orient, and that it was propagated thence to Hellas 
and Italy, while the Romans (according to others, the Greeks) trans- 
planted the vine to Gaul and the banks of the Rhine. 1 For botanical 
reasons, A. de Candolle 2 was inclined to regard the region south of 
the Caucasus as "the central and perhaps the most ancient home of 
the species." In view of the Biblical tradition of Noah planting the 
grape-vine near the Ararat, 8 it is a rather attractive hypothesis to con- 
ceive of Armenia as the country from which the knowledge of the 
grape took its starting-point. 4 However, we must not lose sight of the 
fact that both vine and wine were known in Egypt for at least three or 
four millenniums B.C., 6 and were likewise familiar in Mesopotamia at 
a very early date. This is not the place for a discussion of O. Schrader's 
theory 6 that the name and cultivation of the vine are due to Indo- 
Europeans of anterior Asia; the word for "wine" may well be of Indo- 
European or, more specifically, Armenian origin, but this does not 

1 Cf. the excellent study of G. Curtel, La Vigne et le vin chez les Romains 
(Paris, 1903). See also A. Stummer, Zur Urgeschichte der Rebe und des Weinbaues 
{Mitt. Anthr. Ges. Wien, 1911, pp. 283-296). 

1 Origin of Cultivated Plants, p. 192. 

8 Genesis, ix, 20. 

4 Cf. R. Billiard, La Vigne dans l'antiquite\ p. 31 (Lyon, 1913). This is a well 
illustrated and artistic volume of 560 pages and one of the best monographs on the 
subject. As the French are masters in the art of viticulture, so they have also pro- 
duced the best literature on the science of vine and wine. Of botanical works, 
J.-M. Guillon, Etude generate de la vigne (Paris, 1905), may be recommended. 

8 V. Loret, Flore pharaonique, p. 99. 

8 In Hehn, Kulturpflanzen, pp. 91-95. 


The Grape-Vine 221 

prove that the origin of viticulture itself is traceable to Indo-Europeans. 
The Semitic origin seems to me to be more probable. The Chinese 
received the grape-vine in late historical times from Fergana, an Iranian 
country, as a cultivation entirely unknown in previous epochsj and 
it is therefore sufficient for our purpose to emphasize the fact that 
vine-culture in its entire range was at that time firmly established in 
Western Asia, inclusive of Iran. 

The first knowledge of the cultivated vine (Vitis vinifera) and of wine 
produced from its grapes was likewise obtained by the Chinese through 
the memorable mission of General Can K'ien, when in 128 B.C. he 
travelled through Fergana and Sogdiana on his way to the Yue-£i 
and spent a year in Bactria. As to the people of Fergana (Ta-yuan) , 
he reported, "They have wine made of grapes." The same fact he 
learned regarding the Parthians (An-si). It is further stated in the 
same chapter of the Si ki that the wealthy among the people of Fergana 
stored grape-wine in large quantity up to ten thousand gallons (/ft, a 
dry measure) for a long time, keeping it for several decades without 
risk of deterioration; they were fond of drinking wine in the same 
manner as their horses relished alfalfa. The Chinese envoys took the 
seeds of both plants along to their country, and the Son of Heaven was 
the first to plant alfalfa and the vine in fertile soil; and when envoys 
from abroad arrived at the Court, they beheld extensive cultivations of 
these plants not far from the imperial palace. The introduction of the vine 
is as well authenticated as that of alfalfa. The main point to be noted 
is that the grape, in like manner as alfalfa, and the art of making wine, 
were encountered by the Chinese strictly among peoples of Aryan 
descent, principally of the Iranian family, not, however, among any 
Turkish tribes. 

According to the Han Annals, the kingdom Li-yi ^ -^, which 
depended on Sogdiana, produced grapes; and, as the water of that 
country is excellent, its wine had a particular reputation. 2 

K'ah (Sogdiana) is credited with grapes in the Annals of the Tsin 
Dynasty. 3 Also grape-wine was abundant there, and the rich kept up to 
a thousand gallons of it. 4 The Sogdians relished wine, and were fond of 
songs and dances. 6 Likewise in Si (Tashkend) it was a favorite bever- 

1 This is also the conclusion of J. Hoops (Waldbaume und Kulturpflanzen, 
p. 561). 

2 Hou Han Im, Ch. n8 f p. 6 (cf. Chavannes, Toung Poo, 1907, p. 195). 
* Tsin Su, Ch. 97, p. 6 b (ibid., p. 6: grape- wine in Ta-yuan or Fergana). 
4 Sui $u, Ch. 83, p. 4 b. 

1 Van Su, Ch. 221 B, p. 1. 


age. 1 When the Sogdian K'afi Yen-tien in the first part of the seventh 
century a.d. established a Sogdian colony south of the Lob Nor, he 
founded four new cities, one of which was called "Grape City" (P'u- 
t'ao S'en) ; for the vine was planted in the midst of the town. 2 

The Iranian Ta Yue-Si or Indo-Scythians must also have been in 
possession of the vine, as we are informed by a curious text in the 
Kin lou tse & Wt fF", 3 written by the Emperor Yuan JQ (a.d. 552-555) 
of the Liang dynasty. "The people in the country of the Great Yue-ci 
are clever in making wine from grapes, flowers, and leaves. Sometimes 
they also use roots and vegetable juice, which they cause to ferment. 4 
These flowers resemble those of the clove-tree (ttn-hian T ?F, Caryo- 
phyllus aromaticus) , but are green or bright-blue. At the time of 
spring and summer, the stamens of the flowers are carried away and 
scattered around by the wind like the feathers of the bird Iwan 5K. 
In the eighth month, when the storm blows over the leaves, they are 
so much damaged and torn that they resemble silk rags: hence people 
speak of a grape-storm (p'u-t'aofun), or also call it 'leaves-tearing storm' 
(lie ye fun fSi MM.)." 

Finally we know also that the Aryan people of Kuca, renowned 
for their musical ability, songs, and dances, were admirers of grape- 
wine, some families even storing in their houses up to a thousand hu 
fil of the beverage. This item appears to have been contained in the 
report of General Lu Kwan S it, who set out for the conquest of Kuca 
in a.d. 384. 8 

In the same manner as the Chinese discovered alfalfa in Ki-pin 
(Kashmir), they encountered there also the vine. 6 Further, they found 
it in the countries Tsiu-mo M. ~M 7 and Nan-tou ft tf&. 

1 T'ai p'in hwan yil ki, Ch. 186, p. 7 b; also in Yen-k'i (Karasar): Cou Su, 
Ch. 50, p. 4 b. 

■ Pelliot, Journal asiatique, 1916, I, p. 122. 8 Ch. 5, p. 23. 

4 Strabo (XI. xm, 1 1) states that the inhabitants of the mountainous region 
of northern Media made a wine from some kind of roots. 

5 Other sources fix the date in the year 382 (see Sylvain Levi, Le "Tokharien 
B," langue de Koutcha, Journal asiatique, 1913, II, p. 333). The above fact is 
derived from the Hou lian lu $£ ^ fffc, quoted in the T'ai P'in yu Ian (Ch. 972, p. 3) ; 
see also T'an Su, Ch. 221 A, p. 8. We owe to S. L6vi the proof that the people of 
Ku2a belong to the Indo-European family, and that their language is identical with 
what was hitherto known from the manuscripts discovered in Turkistan as 
Tokharian B. 

6 Ts'ien Han Su, Ch. 96 A, p. 5. Kashmir was still famed for its grapes in the 
days of the Emperor Akbar (H. Blochmann, Ain I Akbari, Vol. I, p. 65), but at 
present viticulture is on the decline there (Watt, Commerical Products of India, 
pp. 1 1 12, 1 1 14). 

r Regarding this name, see Chavannes, Les Pays d'occident d'apre6 le Wei 
lio (T'oung Poo, 1905, p. 536). 

The Grape-Vine 223 

In the T'ang period the Chinese learned also that the people of 
Fu-lin (Syria) relished grape-wine, 1 and that the country of the Arabs 
(Ta-si) produced grapes, the largest of the size of fowl's eggs. 1 In 
other texts such grapes are also ascribed to Persia. 3 At that epoch, 
Turkistan had fallen into the hands of Turkish tribes, who absorbed 
the culture of their Iranian predecessors; and it became known to the 
Chinese that the Uigur had vine and wine. 

Viticulture was in a high state of development in ancient Iran. 
Strabo 4 attributes to Margiana (in the present province of Khorasan) 
vines whose stock it would require two men with outstretched arms to 
clasp, and clusters of grapes two cubits long. Aria, he continues, is 
described as similarly fertile, the wine being still richer, and keeping 
perfectly for three generations in unpitched casks. Bactriana, which 
adjoins Aria, abounds in the same productions, except the olive. 

The ancient Persians were great lovers of wine. The best vintage- 
wines were served at the royal table. 6 The couch of Darius was over- 
shadowed by a golden vine, presented by Pythius, a Lydian. 8 The 
inscription of Persepolis informs us that fifty congius 7 of sweet wine 
and five thousand congius of ordinary wine were daily delivered to the 
royal house. 8 The office of cup-bearer in the palace was one of im- 
portance. 9 The younger Cyrus, when he had wine of a peculiarly fine 
flavor, was in the habit of sending half -emptied flagons of it to some 
of his friends, with a message to this effect: "For some time Cyrus has 
not found a pleasanter wine than this one; and he therefore sends some 
to you, begging you to drink it to-day with those whom you love 
best." 10 

Strabo 11 relates that the produce of Carmania is like that of Persia, 
and that among other productions there is the vine. "The Carmanian 

1 Hirth, China and the Roman Orient, pp. 58, 63. 

* T'ai p'in hwan yii ki, Ch. 1 86, p. 15 b. 

3 For instance, Pen ts'ao yen i, Ch. 18, p. I (ed. of Lu Sin-yuan). 

* II. 1, 14, and XI. x, 2. 

6 Esther, 1, 7 ("And they gave them drink in vessels of gold, the vessels being 
diverse one from another, and royal wine in abundance, according to the state of 
the king"). 

•Herodotus, vn, 27; Athenaeus, xn, 514 f. According to G. W. Elderkin 
(Am. Journal of Archaeology, Vol. XXI, 1917, p. 407), the ultimate source of this 
motive would be Assyrian. 

7 A measure of capacity equal to about six pints. 

8 Joret, Plantes dans l'antiquite, Vol. II, p. 95. 

9 Xenophon, Cyropaedia, I. in, 8-9. 

10 Xenophon, Anabasis, I. IX, 25. 
» XV. 11, 14. 

224 Sino-Iranica 

vine, as we call it, often bears bunches of grapes of two cubits in size, 
the seeds being very numerous and very large; probably the plant 
grows in its native soil with great luxuriance." The kings of Persia were 
not content, however, with wines of native growth; but when Syria 
was united with their empire, the Chalybonian wine of Syria became 
their privileged beverage. 1 This wine, according to Posidonius, was 
made in Damascus, Syria, from vines planted there by the Persians. 2 

Herodotus 8 informs us that the Persians are very fond of wine and 
consume it in large quantities. It is also their custom to discuss im- 
portant affairs in a state of intoxication; and on the following morning 
their decisions are put before them by the master of the house where 
the deliberations have been held. If they approve of the decision in the 
state of sobriety, they act accordingly; if not, they set it aside. When 
sober at their first deliberation, they always reconsider the matter under 
the influence of wine. In a similar manner, Strabo 4 says that their 
consultations on the most important affairs are carried on while drink- 
ing, and that they consider the resolutions made at that time more to 
be depended upon than those made when sober. In the Sahnameh, 
the Persian epic, deliberations axe held during drinking-bouts, but 
decision is postponed till the following day. 6 Cambyses was ill reputed 
for his propensity for wine. 6 Deploring the degeneracy of the Persians, 
Xenophon 7 remarks, "They continue eating and drinking till those 
who sit up latest go to retire. It was a rule among them not to bring 
large cups to their banquets, evidently thinking that abstinence from 
drinking to excess would less impair their bodies and minds. The 
custom of not bringing such vessels still continues; but they drink so 
excessively that instead of bringing in, they are themselves carried out, 
as they are no longer able to walk upright." Procopius, the great 
Byzantine historian of the sixth century, 8 says that of all men the 
Massagetae (an Iranian tribe) are the most intemperate drinkers. So 

1 Strabo, XV. m, 22. 

2 Athenaeus, 1. 

8 I. 133- 

4 XV. in, 20. 

6 F. Spiegel, Eranische Altertumskunde, Vol. Ill, p. 672. Cf . what John Fryer 
(New Account of East India and Persia being Nine Years' Travels 1672-81, Vol. II, 
p. 210, ed. of Hakluyt Society) says of the modern Persians: "It is incredible to see 
what quantities they drink at a merry-meeting, and how unconcerned the next day 
they appear, and brisk about their business, and will quaff you thus a whole week 

6 Herodotus, in, 34. 

7 Cyropaedia, VIII. vni, 9-10. 

8 Historikon, III. xil, 8. 

The Grape-Vine 225 

were also the Sacae, who, maddened with wine, were defeated by 
Cyrus. 1 In the same passage, Strabo speaks of a Bacchanalian festival 
of the Persians, in which men and women, dressed in Scythian style, 
passed day and night in drinking and wanton play. On the other 
hand, it must not be forgotten that such judgments passed by one 
nation on another are usually colored or exaggerated, and must be 
accepted only at a liberal discount; also temperance was preached in 
ancient Persia, and intemperance was severely punished. 2 With all 
the evils of over-indulgence in wine and the social dangers of alcohol, 
the historian, whose duty it is to represent and to interpret phe- 
nomena as they are, must not lose sight of the fact that wine con- 
stitutes a factor of economic, social, and cultural value. It has largely 
contributed to refine and to intensify social customs and to heighten 
sociability, as well as to promote poetry, music, and dancing. It has 
developed into an element of human civilization, which must not 
be underrated. Temperance literature is a fine thing, but who would 
miss the odes of Anakreon, Horace, or Hafiz? 

The word for the grape, brought back by Can K'ien and still current 
in China and Japan (budo) , is W $ii (ancient phonetic spelling of the 
Han Annals, subsequently 36 3tJ) 3 p'u-t'ao, *bu-daw, "grape, vine". Since 
Can K'ien made the acquaintance of the grape in Ta-yuan (Fergana) 
and took its seeds along from there to China, it is certain that he also 
learned the word in Fergana; hence we are compelled to assume that,) 
*bu-daw is Ferganian, and corresponds to an Iranian *budawa or 
*bu5awa, formed with a suffix wa or awa, from a stem buda, which in 
my opinion may be connected with New Persian bdda ("wine") and 
Old Persian fiaTiiicr) ("wine- vessel ") = Middle Persian batak, New 
Persian bddye.* The Sino-Iranian word might also be conceived as a 
dialectic form of Avestan madav ("wine from berries"). 

It is well known that attempts have been made to derive the Chinese 
word from Greek (56rpvs ("a bunch of grapes"). Tomaschek 5 was 
the first to offer this suggestion; T. Kingsmill 8 followed in 1879, and 

1 Strabo, XI. vm, 5. 

2 Cf. Jackson, in Grundriss der iranischen Philologie, Vol. II, p. 679. 

3 The graphic development is the same as in the case'of mu-su (see above, p. 212). 

4 Cf. Horn, Neupersische Etymologie, No. 155. The Chinese are fond of etymol- 
ogizing, and Li Si-cen explains the word p'u-t'ao thus: "When people drink (p'u 
§j§) it, they become intoxicated {Vao §&0)." The joke is not so bad, but it is 
no more than a joke. 

6 Sogdiana, Sitzungsber. Wiener Akad., 1877, p. 133. 

8 Journal China Branch Roy. As. Soc, Vol. XIV, pp. 5, 19. 

226 Sino-Iranica 

Hirth 1 endorsed Kingsmill. No one gave a real demonstration of the 
case. Tomaschek argued that the dissemination of the vine in Central 
Asia is connected with Macedonian-Greek rule and Hellenic influence. 
This is decidedly wrong, for the vine grows spontaneously in all north- 
ern Iranian regions; and its cultivation in Iran is traceable to a great 
antiquity, and is certainly older there than in Greece. The Greeks 
received vine and wine from western Asia. 2 Greek (36rpvs, in all likeli- 
hood, is a Semitic loan-word. 3 It is highly improbable that the people 
of Fergana would have employed a Greek word for the designation of 
a plant which had been cultivated in their dominion for ages, nor is 
there any evidence for the silent admission that Greek was ever known 
or spoken in Fergana at the time of Can K'ien's travels. The influence 
of Greek in the Iranian domain is extremely slight: nothing Greek has 
as yet beeij found in any ancient manuscripts from Turkistan. In 
my opinion, there is no connection between p % u-Vao and fiorpvs, nor 
between the latter and Iranian *budawa. 

It is well known that several species of wild vine occur in China, in 
the Amur region, and Japan. 4 The ancient work Pie In is credited with 
the observation that the vine (p'u-fao) grows in Lun-si (Kan-su) , Wu-yuan 
3l 0. (north of the Ordos), and in Tun-hwan (in Kan-su). 6 Li Si-cen 
therefore argues that in view of this fact the vine must of old have existed 
in Lun-si in pre-Han times, but had not yet advanced into Sen-si. It 
is inconceivable how Bretschneider 6 can say that the introduction of 
the grape by Can K'ien is inconsistent with the notice of the grape in 
the earliest Chinese materia medica. There is, in fact, nothing alarming 
about it: the two are different plants; wild vines are natives of northern 

1 Fremde Einflusse in der chin. Kunst, p. 28; and Journal Am. Or. Soc, 
Vol. XXXVII, 1917, p. 146. Hirth's arguments are based on unproved premises. The 
grape-design on the so-called grape mirrors has nothing to do with Greek or Bactrian 
art, but comes from Iranian-Sasanian art. No grape mirrors were turned out under the 
Han, they originated in the so-called Leu-c'ao period from the fourth to the seventh 
century. The attribution "Han" simply rests on the puerile assumption made in 
the Po ku Vu lu that, because Can K'ien introduced the grape, the artistic designs 
of grapes must also have come along with the same movement. 

1 Only a "sinologue" could assert that the grape was "originally introduced 
from Greece, vid Bactria, about 130 B.C." (Giles, Chinese Dictionary, No. 9497). 

8 Muss-Arnolt, Transactions Am. Phil. Assoc, Vol. XXIII, 1892, p. 142. 
The variants in spelling pdarpvxos, /36rpuxo$, plainly indicata the status of a loan- 
word. In Dioscorides (in, 120) it denotes an altogether different plant, — Chen- 
opodium botrys. 

* The Lo-lo of Yun-nan know a wild grape by the name ko-p'i-ma, with large, 
black, oblong berries (P. Vial, Dictionnaire francais-lolo, p. 276). The grape is 
te-mu-se-ma in Nyi Lo-lo, sa-lu-zo or sa-io-10 in Ahi Lo-lo. 

' Pen ts'ao kan mu, Ch. 33, p. 3. 

8 Bot. Sin., pt. Ill, p. 438. 

The Grape-Vine 227 

China, but have never resulted in a cultivation; the cultivated species 
(Vitis vinifera) was introduced from Iran, and never had any relation 
to the Chinese wild species (Vitis bryoniaefolia) . In a modern work, 
Mun ts'iian tsa yen It M. ft H", 1 which gives an intelligent discussion 
of this question, the conclusion is reached that the species from Fergana 
is certainly different from that indigenous to China. The only singular 
point is that the Pie lu employs the Ferganian word p'u-Vao with refer- 
ence to the native species; but this is not an anachronism, for the Pie lu 
was written in post-Christian times, centuries after Can K'ien; and it 
is most probable that it was only the introduced species which gave the 
impetus to the discovery of the wild species, so that the latter received 
the same name. 2 

Another wild vine is styled yin-yii H JL (Vitis bryoniaefolia or 
V. labrusca), which appears in the writings of T'ao Hun-kin (a.d. 
451-536) and in the T'an pen ts'ao of Su Kun, but this designation has 
reference only to a wild vine of middle and northern China. Yen Si-ku 
(a.d. 579-645), in his K'an miu leh su ^!J H£ IE f&, 3 ironically remarks 
that regarding the yin-yii as a grape is like comparing the U #* (Poncirus 
trijoliata) of northern China with an orange (kii fj|); that the yin-yii, 
although a kind of p'u-t'ao, is widely different from the latter; and that 
the yin-yii of Kian-nan differs again from the yin-yii of northern China. 
Hirth's theory, 4 that this word might represent a transcription of 
New Persian angur, is inadmissible. We have no right to regard Chinese 
words as of foreign origin, unless these are expressly so indicated by the 
Chinese philologists who never fail to call attention to such borrowing. 
If this is not the case, specific and convincing reasons must be adduced 
for the assumption that the word in question cannot be Chinese. There 
is no tradition whatever that would make yin-yii an Iranian or a foreign 
word. The opposite demonstration lacks any sound basis: New Persian, 
which starts its career from the end of the tenth century, could not come 
into question here, but at the best Middle Persian, and angur is a 
strictly New-Persian type. A word like angur would have been dis- 
sected by the Chinese into an-\-gut (gur), but not into an-\-uk; more- 
over, it is erroneous to suppose that final k can transcribe final r; 6 
in Iranian transcriptions, Chinese final k corresponds to Iranian fe, 
g, or the spirant x. It is further inconceivable that the Chinese might 

1 T'u Su tsi Ve*, xx, Ch. 113. 

2 Compare the analogous case of the walnut. 
8 Ch. 8, p. 8 b (ed. of Hu pei Is'uA Su). 

4 Fremde Einflusse in der chinesischen Kunst, p. 17. 
• Compare above, p. 214. 

228 Sino-Iranica 

have applied a Persian word designating the cultivated grape to a 
wild vine which is a native of their country, and which particularly 
grows in the two Kian provinces of eastern China. The Gazetteer of 
Su-cou 1 says expressly that the name for the wild grape, Ian p'u-t'aoy 
in the Kian provinces, is yih-yti. Accordingly it may be an ancient 
term of the language of Wu. The Pen ts'ao kan mu 2 has treated yin-yii 
as a separate item, and Li Si-cen annotates that the meaning of the 
term is unexplained. It seems to me that for the time being we have 
to acquiesce in this verdict. Yen-yii %& J| and yih-le §§ f§" are added 
by him as synonymes, after the Mao H ^ !Nf and the Kwan ya, while 
ye p'u-Vao ("wild grape") is the common colloquial term (also fen 
mih or mu lun HI & /fc ft). It is interesting to note that the earliest 
notices of this plant come only from Su Kuii and C'en Ts'ah-k'i of the 
T'ang dynasty. In other words, it was noted by the Chinese naturalists 
more than seven centuries later than the introduction of the cultivated 
grajpe, — sufficient evidence for the fact that the two are not in any way 

It must not be imagined that with Can K'ien's deed the introduction 
of the vine into China was an accomplished fact; but introductions of 
seeds were subsequently repeated, and new varieties were still imported 
from Turkistan by K'an-hi. There are so many varieties of the grape 
in China, that it is hardly credible that all these should have at once 
been brought over by a single man. It is related in the Han Annals 
that Li Kwan-li ^ M ffl, being General of Er-§i — U (*Ni-S'i), after 
the subjugation of Ta-yuan, obtained grapes which he took along to 

Three varieties of grape are indicated in the Kwan H* written 
before a.d. 527, — yellow, black, and white. The same varieties are 
enumerated in the Yu yan tsa tsu, while Li Si-cen speaks of four varie- 
ties, — a round one, called ts'ao lun tu ^ tl & ("vegetable dragon- 
pearls"); a long one, ma %u p'u-fao (see below); a white one, called 
"crystal grapes" {Swi tsin p'u-t'ao); and a black one, called "purple 
grapes" (tse ^ p'u-Vao), — and assigns to Se-c'wan a green ($&) grape, 
to Yun-nan grapes of the size of a jujube. 4 Su Sun of the Sung mentions 
a variety of seedless grapes. 

1 Su lou fu Si, Ch. 20, p. 7 b. 

2 Ch. 33, p. 4. 

8 T'ai p'in yii Ian, Ch. 972, p. 3. 

4 T'an Ts'ui fU ^ , in his valuable description of Yun-nan (Tien hat yu 
hen Si, published in 1799, Ch. 10, p. 2, ed. of Wen yin lou yii ti ts'un Su), states that the 
grapes of southern Yun-nan are excellent, but that they cannot be dried or sent to dis- 
tant places. 

The Grape-Vine 229 

In Hah-cou yellow and bright white grapes were styled Zu-tse & •? 
("beads, pearls"); another kind, styled "rock-crystal" {swi-tsin), ex- 
:elled in sweetness; those of purple and agate color ripened at a little 
ater date. 1 

To Turkistan a special variety is attributed under the name so-so 
85 3§t grape, as large as wu-wei-tse 3l B£ J- ("five flavors," Schizandra 
:hinensis) and without kernels M$%. A lengthy dissertation on this 
xuit is inserted in the Pen ts'ao kan mu Si i* The essential points are 
she following. It is produced in Turf an and traded to Peking; in appear- 
ance it is like a pepper-corn, and represents a distinct variety of grape, 
[ts color is purple. According to the Wu tsa tsu 3l jJS fi, written in 
16 10, when eaten by infants, it is capable of neutralizing the poison of 
;mall-pox. The name so-so is not the reproduction of a foreign word, 
Dut simply means "small." This is expressly stated in the Pen kin fun 
Viian ^ $£ %. W> , which says that the so-so grapes resemble ordinary 
grapes, but are smaller and finer, and hence are so called (U0 W. &H 
§fc &). The Pi e'en ¥ M of Yu-wen Tin ^ 3C % annotates, however, 
that so-so is an error for sa-so IS^, without giving reasons for this 
opinion. Sa-so was the name of a palace of the Han emperors, and this 
substitution is surely fantastic. Whether so-so really is a vine-grape 
seems doubtful. It is said that so-so are planted everywhere in China 
to be dried and marketed, being called in Kiah-nan/aw p'u-Vao ("foreign 
*rape"). 8 

The Emperor K'an-hi (1662-1722), who knew very well that grapes 
bad come to China from the west, tells that he caused three new varie- 
ties to be introduced into his country from Hami and adjoining terri- 
tories, — one red or greenish, and long like mare-nipples; one not very 
large, but of agreeable taste and aroma; and another not larger than a 
pea, the most delicate, aromatic, and sweetest kind. These three varie- 
ties of grape degenerate in the southern provinces, where they lose 
their aroma. They persist fairly well in the north, provided they are 
planted in a dry and stony soil. "I would procure for my subjects," 
the Emperor concludes, "a novel kind of fruit or grain, rather than 
build a hundred porcelain kilns." 4 

Turkistan is well known to the Chinese as producing many varieties 

1 Mon lian /« |£ He £&> by WuTse-mu ^ g $ of the Sung (Ch. 18, p. 5 b; 
id. of Ci pu tsu lai ts'un Su). 

1 Ch. 7, p. 69. This valuable supplement to the Pen ts'ao kan mu was first 
published in 1650 (reprinted 1765 and appended to several modern editions of the 
Pen ts'ao) by Cao Hio-min |§ ^ fffc (hao §u-hien $3 $f ) of Hah-cou. 

8 Mun ts'iian tsa yen ^| ^. f§ H\ cited in T'u Su tsi I'en, XX, Ch. 130. 

4 Memoires concernant les Chinois, Vol. IV, 1779, pp. 471-472. 

t$o Sino-Iranica 

of grape. According to the Hut k'ian li 5H JS ("Records of Turkis- 
tan"), written in 1772 by the two Manchu officers Fusambd and Surde, 
"there are purple, white, blue, and black varieties; further, round and 
long, large and small, sour and sweet ones. There is a green and seed- 
less variety, comparable to a soy-bean, but somewhat larger, and of 
very sweet and agreeable flavor [then the so-so is mentioned]. Another 
kind is black and more than an inch long; another is white and large. 
All varieties ripen in the seventh or eighth month, when they are 
dried and can be transported to distant places." According to the 
Wu tsa tsu, previously quoted, Turkistan has a seedless variety of 
grape, called tu yen 3& BR p'u-Vao ("hare-eye grape"). 

A. v. Le Coq 1 mentions under the name sozuq saivi a cylindrical, 
whitish-yellow grape, the best from Toyoq and Bulayiq, red ones of 
the same shape from Manas and ShichO. Sir Aurel Stein 1 says that 
throughout Chinese Turkistan the vines are trained along low fences, 
ranged in parallel rows, and that the dried grapes and currants of 
Ujat find their way as far as the markets of Aksu, Kashgar, and Turfan. 

Every one who has resided in Peking knows that it is possible to 
obtain there during the summer seemingly fresh grapes, preserved from 
the crop of the previous autumn, and that the Chinese have a method of 
preserving them. The late F. H. King, 3 whose studies of the agriculture 
of China belong to the very best we have, observed regarding this 
point, " These old people have acquired the skill and practice of storing 
and preserving such perishable fruits as pears and grapes so as to 
enable them to keep them on the market almost continuously. Pears 
were very common in the latter part of June, and Consul-General 
Williams informed me that grapes are regularly carried into July. In 
talking with my interpreter as to the methods employed, I could only 
learn that the growers depend simply upon dry earth cellars which can 
be maintained at a very uniform temperature, the separate fruits being 
wrapped in paper. No foreigner with whom we talked knew their 
methods." This method is described in the Ts'i min yao $u, an ancient 
work on husbandry, probably from the beginning of the sixth century, 4 
although teeming with interpolations. A large pit is dug in a room of 
the farmhouse for storing the grapes, and holes are bored in the walls 
near the surface of the ground and stuffed with branches. Some of 
these holes are filled with mud to secure proper support for the room. 

1 Sprichworter und Lieder aus Turfan, p. 92. 
■ Sand-Buried Ruins of Khotan, p. 228. 

• Farmers of Forty Centuries, p. 343 (Madison, Wis., 191 1). 

* See Bretschneider, Bot. Sin., pt. I, p. 77; Hirth, T'oung Poo, 1895, p. 436; 
Pelliot, Bulletin de I'Ecolefrangaise, Vol. IX, p. 434. 

The Grape-Vine 231 

The pit in which the grapes are stored is covered with loam, and thus 
an even temperature is secured throughout the winter. 1 

The Jesuit missionaries of the eighteenth century praise the raisins 
of Hoai-lai-hien 1 on account of their size: "Nous parlons d'apres le 
temoignage de nos yeux: les grains de ces grappes de raisins sont gros 
comme des prunes damas-violet, et la grappe longue et grande a propor- 
tion. Le climat peut y faire; mais si les livres disent vrai, cela vient 
originairement de ce qu'on a ente* des vignes sur des jujubiers; et 
I'epaisseur de la peau de ces raisins nous le ferart croire." 3 

Raisins are first mentioned as being abundant in Yun-nan in the 
Y tin-nan ki* ("Memoirs regarding Yun-nan"), a work written in the 
beginning of the ninth century. Li Si-cen remarks that raisins are made 
by the people of the West as well as in T'ai-yuan and P'ih-yan in San-si 
Province, whence they are traded to all parts of China. Hami in 
Turkistan sends large quantities of raisins to Peking. 6 In certain parts 
of northern China the Turkish word Hfrra'i for a small kind of raisin 
is known. It is obtained from a green, seedless variety, said to originate 
from Bokhara, whence it was long ago transplanted to Yarkand. 
After the subjugation of Turkistan under K'ien-luh, it was brought to 
Jehol, and is still cultivated there. 8 

Although the Chinese eagerly seized the grape at the first oppor- 
tunity offered to them, they were slow in accepting the Iranian custom 
of making and drinking wine. 7 The Arabic merchant Soleiman (or 
whoever may be responsible for this account), writing in a.d. 851, 
reports that "the wine taken by the Chinese is made from rice; they 
do not make wine from grapes, nor is it brought to them from abroad; 

1 A similar contrivance for the storage of oranges is described in the Memoires 
concernant les Chinois, Vol. IV, p. 489. 

* I presume that Hwai (or Hwo)-lu hien in the prefecture of Cen-tin, Ci-li 
Province, is meant. 

* Memoires concernant les Chinois, Vol. Ill, 1778, p. 498. 

* T'ai p'in yii Ian, Ch. 972, p. 3. 

1 An article on Hami raisins is inserted in the Memoires concernant les Chinois 
(Vol. V, 1780, pp. 481-486). The introduction to this article is rather strange, an 
effort being made to prove that grapes have been known in China since times of 
earliest antiquity; this is due to a confusion of the wild and the cultivated vine. 
In Vol. II, p. 423, of the same collection, it is correctly stated that vine and wine be- 
came known under the reign of the Emperor Wu. 

6 Cf. O. Franke, Beschreibung des Jehol-Gebietes, p. 76. 

7 The statement that Can K'ien taught his countrymen the art of making wine, 
as asserted by Giles (Biographical Dictionary, p. 12) and L. Wieger (Textes 
historiques, p. 499), is erroneous. There is nothing to this effect in the Si ki or in 
the Han Annals. 

232 Sino-Iranica 

they do not know it, accordingly, and make no use of it." 1 This doubt- 
less was correct for southern China, where the information of the 
Arabic navigators was gathered. The grape, however, is chiefly to be 
found in northern China, 2 and at the time of Soleiman the manu- 
facture of grape-wine was known in the north. The principal document 
bearing on this subject is extant in the history of the T'ang dynasty. 

In a.d. 647 a peculiar variety of grapes, styled ma %u p % u fao $1 
^L li) 3sf ("mare-nipple grapes") were sent to the Emperor T'ai Tsun 
^C ^ by the (Turkish) country of the Yabgu M IS. It was a bunch 
of grapes two feet long, of purple color. 3 On the same occasion it is 
stated, "Wine is used in the Western Countries, and under the 
former dynasties it was sometimes sent as tribute, but only after 
the destruction of Kao-5'an M H (Turf an), when 'mare-nipple grapes' 
cultivated in orchards were received, also the method of making wine 
was simultaneously introduced into China (a.d. 640). T'ai Tsun 
experienced both its injurious and beneficial effects. Grape-wine, when 
ready, shines in all colors, is fragrant, very fiery, and tastes like the 
finest oil. The Emperor bestowed it on his officials, and then for the 
first time they had a taste of it in the capital." 4 

These former tributes of wine are alluded to in a verse of the poet 
Li Po oi the eighth century, "The Hu people annually offered grape- 
wine." 5 Si Wan Mu, according to the Han Wu ti net Iwan of the 
third century or later, is said to have presented grape-wine to the Han 
Emperor Wu, which certainly is an unhistorical and retrospective 

A certain Can Hun-mao DJt ^ M, a native of Tun-hwan in Kan-su, 
is said to have devoted to grape-wine a poem of distinct quality. 6 
The locality Tun-hwan is of significance, for it was situated on the 

1 M. Reinaud, Relation des voyages faits par les Arabes et les Persans dans 
l'lnde et a la Chine, Vol. I, p. 23. \ 

2 In the south, I am under the impression it is rather isolated. It occurs, for 
instance, in Sah-se cou Jh. f§> '}\i in the prefecture of T'ai-p'ih, Kwan-si Province, 
in three varieties, — green, purple, and crystal, — together with an uneatable wild 
grape (San se lou li, Ch. 14, p. 8, ed. published in 1835). "Grapes in the neighbor- 
hood of Canton are often unsuccessful, the alternations of dry heat and rain being 
too much in excess, while occasional typhoons tear the vines to pieces" (J. F. Davis, 
China, Vol. II, p. 305). They occur in places of Fu-kien and in the Chusan Archi- 
pelago (cf. T'u Su tsi Ven, VI, Ch. 1041). 

3 T'an hui yao, Ch. 200, p. 14; also Fun H wen Men ki H .0Q P9 JL fS ( Ch. 7, 
p. 1 b (ed. of Ki fu ts'un Su), by Fun Yen Jhj" Q| of the T'ang. 

* Ibid., p. 15. 

5 Pen ts'ao yen i, Ch. 18, p. I. 

6 This is quoted from the TsHen Han lu "ftj ^ §jfc, a work of the Tsin dynasty, 
in the Si leu kwo Vun ts'iu (T'ai pHn yu Ian, Ch. 972, p. 1 b). 

The Grape- Vine 233 

road to Turkistan, and was the centre from which Iranian ideas radiated 
into China. 

The curious point is that the Chinese, while they received the grape 
in the era of the Han from an Iranian nation, and observed the habit 
of wine-drinking among Iranians at large, acquired the art of wine- 
making as late as the T'ang from a Turkish tribe of Turkistan. The 
Turks of the Han period knew nothing of grapes or wine, quite natu- 
rally, as they were then restricted to what is now Mongolia, where soil 
and climatic conditions exclude this plant. Vine-growing, as a matter 
of course, is compatible solely with a sedentary mode of life; and only 
after settling in Turkistan, where they usurped the heritage of their 
Iranian predecessors, 1 did the Turks become acquainted with grape 
and wine as a gift of Iranians. The Turkish word for the grape, Uigur 
bziim (other dialects iiziim), proves nothing along the line of historical 
facts, as speculated by Vambery. 2 It is even doubtful whether the word 
in question originally had the meaning "grape"; on the contrary, it 
merely seems to have signified any berry, as it still refers to the berries 
and seeds of various plants. The Turks were simply epigones and 
usurpers, and added nothing new to the business of vine-culture. 

In accordance with the introduction of the manufacture of grape- 
wine into China, we find this product duly noted in the Pen ts'ao of 
the T'ang, 3 published about the middle of the seventh century; further, 
in the $i liao pen ts'ao by Moh Sen j§L %9c (second half of the seventh 
century), and in the Pen ts'ao H i by C'en Ts'ah-k'i ffl. W. H, who wrote 
in the K'ai-yuan period (713-741). The T'ah pen ts'ao also refers to 
the manufacture of vinegar from grapes. 4 The Pen ts'ao yen i, pub- 
lished in 1 1 16, likewise enumerates grape-wine among the numerous 
brands of alcoholic beverages. 

The Lian se kun tse ki by Can Yue (667-730) 5 contains an anecdote 
to the effect that Kao-£'ah offered to the Court frozen wine made from 
dried raisins, on which Mr. Kie made this comment: "The taste of 
grapes with thin shells is excellent, while grapes with thick shells are 
bitter of taste. They are congealed in the Valley of Eight Winds 
(Pa fun ku A %> ^) . This wine does not spoil in the course of years." 6 

1 This was an accomplished fact by the end of the fourth century a.d. 

2 Primitive Cultur des turko-tatarischen Volkes, p. 218. 

* Cen lei pen ts'ao, Ch. 23, p. 7. 

* Ibid., Ch. 26, p. 1 b. 

6 See The Diamond, this volume, p. 6. 

8 Pen ts'ao kan mu, Ch. 25, p. 14 b. A different version of this story is quoted 
in the T'ai P'in yii Ian (Ch. 845, p. 6 b). 

«34 Sino-Iranica 

A recipe for making grape-wine is contained in the Pei San tsiu kin 
At tfj 'M ffli, 1 a work on the different kinds of wine, written early in the 
twelfth century by Cu Yi-cuh ^cl^, known as Ta-yin Wen % B§ H. 
Sour rice is placed in an earthen vessel and steamed. Five ounces of 
apricot-kernels (after removing the shells) and two catties of grapes 
(after being washed and dried, and seeds and shells removed) are put 
together in a bowl of thin clay (Sa p'en ty ^£), 2 pounded, and strained. 
Three pecks of a cooked broth are poured over the rice, which is placed 
on a table, leaven being added to it. This mass, I suppose, is used to 
cause the grape-juice to ferment, but the description is too abrupt and 
by no means clear. So much seems certain that the question is of a 
rather crude process of fermentation, but not of distillation (see below). 

Sii T'in # HI, who lived under the Emperor Li Tsun (1224-63) of 
the Southern Sung, went as ambassador to the Court of the Mongol 
Emperor Ogotai (1229-45). His memoranda, which represent the 
earliest account we possess of Mongol customs and manners, were 
edited by P'en Ta-ya ^ ^C #§ of the Sung under the title Hei Ta H Ho 
J& H ♦ B& ("Outline of the Affairs of the Black Tatars"), and pub- 
lished in 1908 by Li Wen-t'ien and Hu Se in the Wen yin lou yii ti is'un 
Su. z Su T'in informs us that grape-wine put in glass bottles and sent 
as tribute from Mohammedan countries figured at the headquarters 
of the Mongol Khan; one bottle contained about ten small cups, and 
the color of the beverage resembled the juice of the Diospyros kaki 
[known in this country as Japanese persimmons] of southern China. 
It was accordingly a kind of claret. The Chinese envoy was told that 
excessive indulgence in it might result in intoxication. 

1 Ch. c, p. 19 b (ed. of Ci pu tsu lai ts'uH Su). The work is noted by Wylie 
(Notes on Chinese Literature, p. 150). 

J Literally, "sand-pot." This is a kind of thin pottery (colloquially called Sa 
kwo ffi §&) peculiar to China, and turned out at Hwai-lu (Ci-li), P'in-tift Sou and 
Lu-nan (San-si), and Yao-cou (Sen-si). Made of clay and sand with an admixture 
of coal-dust, so that its appearance presents a glossy black, it is extremely light 
and fragile; but, on account of their thin walls, water may be heated in these pots 
with a very small quantity of fuel. They are a money and time saving device, and 
hence in great demand among the poor, who depend upon straw and dried grass for 
their kitchen fire. With careful handling, such pots and pans may endure a long 
time. The proverb runs, "The sand-pot will last a generation if you do not hit it"; 
and there is another popular saying, "You may pound garlic in a sand-pan, but you 
can do so but once" (A. H. Smith, Proverbs and Common Sayings from the Chinese, 
p. 204). Specimens of this ware from Yao-cou may be seen in the Field Museum, 
others from Hwai-lu are in the American Museum of New York (likewise collected 
by the writer). The above text of the Sung period is the first thus ,far found by me 
which contains an allusion to this pottery. 

1 This important work has not yet attracted the attention of our science. I hope 
to be able to publish a complete translation of it in the future. 

The GfiAPE-ViNE 235 

In his interesting notice "Le Nom turc du vin dans Odoric de 
Pordenone," 1 P. Pelliot has called attention to the word bor as a 
Turkish designation of grape-wine, adding also that this word occurs 
in a Mongol letter found in Turf an and dated 1398.* I can furnish 
additional proof for the fact that bor is an old Mongol word in the 
sense of wine, although, of course, it may have been borrowed from 
Turkish. In the Mongol version of the epic romance of Geser or Gesar 
Khan we find an enumeration of eight names of liquor, all supposed 
to be magically distilled from araki ("arrack, brandy"). These are: 
aradsa (araja), xoradsa or xuradsa, Siradsa, boradsa, takpa, tikpa, 
marba, tnirba. 1 These terms have never been studied, and, with the 
exception of the first and third, are not even listed in Kovalevski's and 
Golstuntki's Mongol Dictionaries. The four last words are characterized 
as Tibetan by the Tibetan suffix pa or ba. Marwa (corresponding in 
meaning to Tibetan Pah) is well known as a word generally used 
throughout Sikkim and other Himalayan regions for an alcoholic 
beverage. 4 As to tikpa, it seems to be formed after the model of Tibetan 
tig-Pan, the liquor for settling {tig) the marriage-affair, presented by the 
future bridegroom to the parents of his intended. 5 

The terms aradsa, xoradsa or xuradsa, Uradsa, and boradsa, are all 
provided with the same ending. The first is given by Kovalevski 6 
with the meaning "very strong koumiss, spirit of wine." A parallel is 
offered by Manchu in artan ("a liquor prepared from milk"), while 
Manchu arjan denotes any alcoholic drink. The term xoradsa or xuradsa 
may be derived from Mongol xuru-t (-t being suffix of the plural), 
corresponding to Manchu kuru, which designates "a kind of cheese 
made from fermented mare's milk, or cheese prepared from cow's or 
mare's milk with the addition of sugar and sometimes pressed into 
forms." The word Hradsa has been adopted by Schmidt and Kovalevski 
in their respective dictionaries as "wine distilled for the fourth time" 
or "esprit de vin quadruple;" but these explanations are simply based 
on the above passage of Geser, in which one drink is supposed to be 

1 Toung Pao, 1914, pp. 448-453. 

1 Ramstedt's tentative rendering of this word by "beaver" is a double error: 
first, the beaver does not occur in Mongolia and is unknown to the Mongols, its 
easternmost boundary is formed by the Yenisei; second, bor as an animal-name 
means "an otter cub," and otter and beaver are entirely distinct creatures. 

* Text, ed. I. J. Schmidt, p. 65; translation, p. 99. Schmidt transcribes arasa, 
chorasa, etc., but the palatal sibilant is preferable. 

4 Cf. H. H. Risley, Gazetteer of Sikkim, p. 75, where also the preparation is 

1 JXschke, Tibetan Dictionary, p. 364. 

• Dictionnaire mongol, p. 143. 

236 Sino-Iranica 

distilled from the other. This process, of course, is purely fantastic, 
and described as a magical feat; there is no reality underlying it. 

The word boradsa, in my opinion, is derived from the Turkish word 
bor discussed by Pelliot; there is no Mongol word from which it could 
be explained. In this connection, the early Chinese account given 
above of foreign grape-wine among the Mongols gains a renewed 
significance. Naturally it was a rare article in Mongolia, and for this 
reason we hear but little about it. Likewise in Tibet grape-wine is 
scarcely used, being restricted to religious offerings in the temples. 1 

The text of the Geser Romance referred to is also important from 
another point of view. It contains the loan-word ariki, from Arabic 
'araq, which appears in eastern Asia as late as the Mongol epoch 
(below, p. 237). Consequently our work has experienced the influence 
of this period, which is visible also in other instances. 2 The foundation 
of the present recension, first printed at Peking in 17 16, is indeed trace- 
able to the thirteenth and fourteenth centuries; many legends and 
motives, of course, are of a much older date. 

Marco Polo relates in regard to T'ai-yuan fu, called by him Taianfu, 
the capital of San-si Province, "There grow here many excellent vines, 
supplying a great plenty of wine; and in all Cathay this is the only place 
where wine is produced. It is carried hence all over the country." 3 
Marco Polo is upheld by contemporary Chinese writers. Grape-wine 
is mentioned in the Statutes of the Yuan Dynasty. 4 The Yin $an ten 
yao ffc H IE H£, written in 133 1 (in 3 chapters) by Ho Se-hwi ^P $? W, 
contains this account: 5 "There are numerous brands of wine: that 
coming from Qara-Khoja (Ha-la-hwo &o B#J ^C) 6 is very strong, that 
coming from Tibet ranks next. Also the wines from P'ih-yah and T'ai- 

1 Cf. T'oung Pao, 1914, p. 412. 

2 Cf. ibid., 1908, p. 436. 

8 Yule and Cordier, The Book of Ser Marco Polo, Vol. II, p. 13. Klaproth 
(cf. Yule's notes, ibid., p. 16) was quite right in saying that the wine of that locality- 
was celebrated in the days of the T'ang dynasty, and used to be sent in tribute to the 
emperors. Under the Mongols the use of this wine spread greatly. The founder of 
the Ming accepted the offering of wine from T'ai-yuan in 1373, but prohibited its 
being presented again. This fact is contained in the Ming Annals (cf. L. Wieger, 
Textes historiques, p. 201 1). 

4 Yuan tien Ian jq -fill jf£, Ch. 22, p. 65 (ed. 1908). 

6 Pen ts'ao kan mu, Ch. 25, p. 14 b. Regarding that work, cf. the Imperial 
Catalogue, Ch. 116, p. 27 b. 

6 Regarding this name and its history see Pelliot, Journal asiatique, 1912, I, 
p. 582. Qara-Khoja was celebrated for its abundance of grapes (Bretschneider, 
Mediaeval Researches, Vol. I, p. 65). J. Dudgeon (The Beverages of the Chinese, 
p. 27), misreading the name Ha-so-hwo, took it for the designation of a sort of wine. 
Stuart (Chinese Materia Medica, p. 459) mistakes it for a transliteration of "hoi- 

The Grape-Vine 237 

yuan (in San-si) take the second rank. According to some statements, 
grapes, when stored for a long time, will develop into wine through a 
natural process. This wine is fragrant, sweet, and exceedingly strong: 
this is the genuine grape-wine." 1 The Ts'ao mu tse ^ yfc ■?*, written 
in 1378 by Ye Tse-k'i M ~$* ^f, contains the following information: 
"Under the Yuan dynasty grape-wine was manufactured in Ki-nin 
M 3f£ and other circuits S& of San-si Province. In the eighth month 
they went to the T'ai-han Mountain j&ff ]U 2 in order to test the 
genuine and adulterated brands: the genuine kind when water is 
poured on it, will float; the adulterated sort, when thus treated, will 
freeze. 3 In wine which has long been stored, there is a certain portion 
which even in extreme cold will never freeze, while all the remainder is 
frozen: this is the spirit and fluid secretion of wine. 4 If this is drunk, 
the essence will penetrate into a man's arm-pits ffli , and he will die. 
Wine kept for two or three years develops great poison." 

The first author who offers a coherent notice and intelligent discus- 
sion of the subject of grape-wine is Li Si-Sen at the end of the sixteenth 
century. 5 He is well acquainted with the fact that this kind of wine was 
anciently made only in the Western Countries, and that the method of 
manufacturing it was but introduced under, the T'ang after the sub- 
jugation of Kao-c'an. He discriminates between two types of grape- 
wine, — the fermented 18 $ :# , of excellent taste, made from grape- 
juice with the addition of leaven in the same fashion as the ordinary 
native rice-wine (or, if no juice is available, dried raisins may be used), 
and the distilled $& ?@. In the latter method "ten catties of grapes are 
taken with an equal quantity of great leaven (distillers' grains) and 
subjected to a process of fermentation. The whole is then placed in an 
earthen kettle and steamed. The drops are received in a vessel, and 
this liquid is of red color, and very pleasing." There is one question, 
however, left open by Li Si-cen. In a preceding notice on distillation 
£k M he states that this is not an ancient method, but was practised 
only from the Yuan period; he then describes it in its application to rice- 
lands," or maybe "alcohol." The latter word has never penetrated into China in 
any form. Chinese a-la-ki does not represent the word "alcohol," as conceived by 
some authors, for instance, J. Macgowan (Journal China Branch Roy. As. Soc, 
Vol. VII, 1873, p. 237); see the following note. 

1 This work is also the first that contains the word a-la-ki fSj jf§ "f$, from 
Arabic 'araq (see T'oung Pao, 1916, p. 483). 

2 A range of mountains separating §an-si from Ci-li and Ho-nan. 

8 This is probably a fantasy. We can make nothing of it, as it is not stated how 
the adulterated wine was made. 

4 This possibly is the earliest Chinese allusion to alcohol. 
6 Pen ts'ao kan mu, Ch. 25, p. 14 b. 

«38 Sino-Iranica 

wine in the same manner as for grape-wine. Certain it is that distillation 
is a Western invention, and was unknown to the ancient Chinese. 1 
Li &i-5en fails to inform us as to the time when the distillation of grape- 
wine came into existence. If this process had become known in China 
under the T'ang in connection with grape-wine, it would be strange if 
the Chinese did not then apply it to their native spirits, but should have 
waited for another foreign impulse until the Mongol period. On the 
other hand, if the method due to the Uigur under the T'ang merely 
applied to fermented grape-wine, we may justly wonder that the Chinese 
had to learn such a simple affair from the Uigur, while centuries earlier 
they must have had occasion to observe this process among many 
Iranian peoples. It would therefore be of great interest to seize upon 
a document that would tell us more in detail what this method of 
manufacture was, to which the T'ang history obviously attaches so 
great importance. It is not very likely that distillation was involved; 
for it is now generally conceded that the Arabs possessed no knowledge 
of alcohol, and that distillation is not mentioned in any relevant litera- 
ture of the Arabs and Persians from the tenth to the thirteenth cen- 
tury. 1 The statement of Li Si-cen, that distillation was first practised 
under the Mongols, is historically logical and in keeping with our 
present knowledge of the subject. It is hence reasonable to hold (at 
least for the present) also that distilled grape-wine was not made 
earlier in China than in the epoch of the Yuan. Mon Sen of the T'ang 
says advisedly that grapes can be fermented into wine, and the recipe 
of the Sung does not allude to distillation. 

In the eighteenth century European wine also reached China. A 
chest of grape-wine figures among the presents made to the Emperor 
K'ah-hi on the occasion of his sixtieth birthday in 171 5 by the Jesuits 
Bernard Kilian Stumpf, Joseph Suarez, Joachim Bouvet, and Domini- 
cus Parrenin. 8 

P. Osbeck, 4 the pupil of Linn6, has the following notice on the 
importation of European wine into China: "The Chinese wine, which 
our East India traders call Mandarin wine, is squeezed out of a fruit 
which is here called Pausio,* and reckoned the same with our grapes. 

*Cf. Bretschneider, Bot. Sin., pt. II, p. 155; J. Dudgeon, The Beverages of 
the Chinese, pp. 19-20; Edkins, China Review, Vol. VI, p. 211. The process of 
distillation is described by H. B. Gruppy, Samshu-Brewing in North China {Journal 
China Branch Roy. As. Soc, Vol. XVIII, 1884, pp. 163-164). 

1 E. O. v. Lippmann, Abhandlungen, Vol. II, pp. 206-209; cf. also my remarks 
in American Anthropologist, 191 7, p. 75. 

1 Cf. Wan Sou Sen tien f$ H #£ J&, Ch. 56, p. 12. 

4 A Voyage to China and the East Indies, Vol. I, p. 315 (London, 1771). 

* Apparently a bad or misprinted reproduction of P'u-t'ao. 

The Grape-Vine 239 

This wine was so disagreeable to us, that none of us would drink it. 
The East India ships never fail taking wine to China, where they often 
sell it to considerable advantage. The Xeres (sherry) wine, for which 
at Cadiz we paid thirteen piastres an anchor, we sold here at thirty- 
three piastres an anchor. But in this case you stand a chance of having 
your tons split by the heat during the voyage. I have since been told, 
that in 1754, the price of wine was so much lowered at Canton, that 
our people could with difficulty reimburse themselves. The Spaniards 
send wines to Manilla and Macao, whence the Chinese fetch a con- 
siderable quantity, especially for the court of Peking. The wine of 
Xeres is more agreeable here than any other sort, on account of its 
strength, and because it is not liable to change by heat. The Chinese 
are very temperate in regard to wine, and many dare not empty a single 
glass, at least not at once. Some, however, have learned from foreigners 
to exceed the limits of temperance, especially when they drink with 
them at free cost." 

Grape-wine is attributed by the Chinese to the Arabs. 1 The 
Arabs cultivated the vine and made wine in the pre-Islamic epoch. 
Good information on this subject is given by G. Jacob. 1 

Theophrastus 8 states that in India only the mountain-country has 
the vine and the olive. Apparently he hints at a wild vine, as does also 
Strabo, 4 who says after Aristobulus that in the country of Musicanus 
(Sindh) there grows spontaneously grain resembling wheat, and a vine 
producing wine, whereas other authors affirm that there is no wine in 
India. Again, he states 5 that on the mountain Meron near the city 
Nysa, founded by Bacchus, there grows a vine which does not ripen 
its fruit; for, in consequence of excessive rains, the grapes drop before 
arriving at maturity. They say also that the Sydracae or Oxydracae 
are descendants of Bacchus, because the vine grows in their country. 
The element -dracae (drakai) is probably connected with Sanskrit 
drdk?a ("grape"). These data of the ancients are vague, and do not 
prove at all that the grape-vine has been cultivated in India from time 
immemorial, as inferred by Joret. 8 Geographically they only refer to 
the regions bordering on Iran. The ancient Chinese knew only of grapes 
in Kashmir (above, p. 222). The Wei £w 7 states that grapes were ex- 

1 Hirth, Chao Ju-kua, pp. 115, 121. 

2 Altarabisches Beduinenleben, 2d ed., pp. 96-109. 
8 Hist, plant., IV. IV, 1 1. 

4 xv, 22. 

• XV. i, 8. 

• Plantes dans l'antiquite, Vol. II, p. 280. 
■ Ch. 102, p. 8. 

240 Sino-Iranica 

ported from Pa-lai $L fft (*Bwat-lai) in southern India. Huan Tsan 1 
enumerates grapes together with pears, crab-apples, peaches, and 
apricots, 2 as the fruits which, from Kashmir on, are planted here and 
there in India. The grape, accordingly, was by no means common in 
India in his time (seventh century). 

The grape is not mentioned in Vedic literature, and Sanskrit draksa 
I regard with Spiegel 3 as a loan-word. Viticulture never was extensive 
or of any importance in Indian agriculture. Prior to the Moham- 
medan conquest, we have little precise knowledge of the cultivation of 
the vine, which was much fostered by Akbar. In modern times it is 
only in Kashmir that it has been received with some measure of 

Huan Tsan 4 states that there are several brands of alcoholic and 
non-alcoholic beverages in India, differing according to the castes. 
The Ksatriya indulge in grape and sugar-cane wine. The Vaigya take 
rich wines fermented with yeast. The Buddhists and Brahmans partake 
of a syrup of grapes or sugar-cane, which does not share the nature 
of any wine. 5 In Jataka No. 183, grape-juice (muddikdpanam) of in- 
toxicating properties is mentioned. 

Huan Yin 6 gives three Sanskrit words for various kinds of wine: — 

(1) 2<l II su-lo, *su5-la, Sanskrit surd, explained as rice-wine 
MM. 7 

1 Ta T'an si yii ki, Ch. 2, p. 8. 

2 Not almond-tree, as erroneously translated by Julien (Memoires, Vol. I, 
p. 92). Regarding peach and apricot, see below, p. 539. 

3 Arische Periode, p. 41. 

4 Ta T'an si yii ki, Ch. 2, p. 8 b. 

6 S. Julien (Memoires, Vol. I, p. 93) translates wrongly, "qui different tout a 
fait du vin distilleV' Distilled wine was then unknown both to the Chinese and in 
India, and the term is not in the text. "Distillation of wines" is surely not spoken 
of in the Cukranlti, as conceived by B. K. Sakkar (The Sukraniti, p. 157; and Hindu 
Sociology, p. 166). 

6 Yi Is'ie kin yin i, Ch. 24, p. 8 b. 

7 This definition is of some importance, for in Boehtlingk's Sanskrit Dictionary 
the word is explained as meaning "a kind of beer in ancient times, subsequently, 
however, in most cases brandy," which is certainly wrong. Thus also O. Schrader's 
speculation (Sprachvergleichung, Vol. II, p. 256), connecting Finno-Ugrian sara, 
sur, etc. ("beer") with this word, necessarily falls to the ground. Macdonell and 
Keith (Vedic Index, Vol. II, p. 458) admit that "the exact nature of sura is not 
certain, it may have been a strong spirit prepared from fermented grains and plants, 
as Eggeling holds, or, as Whitney thought, a kind of beer or ale." It follows also 
from Jataka No. 512 that sura was prepared from rice. In Cosmas' Christian 
Topography (p. 362, ed. of Hakluyt Society) we have fayxooobpa ("coconut- 
wine"); here sura means "wine," while the first element may be connected with 
Arabic ranej or ranj ("coco-nut"). 

The Grape-Vine 241 

(2) j3§ M %$ mi-li-ye, *mei-li(ri)-ya, answering to Sanskrit maireya, 
explained as a wine mixed from roots, stems, flowers, and leaves. 1 

(3) "M $£ mo-Vo, *mwa5-do, Sanskrit madhu, explained as "grape- 
wine" (p'u-t'ao tsiu). The latter word, as is well known, is connected 
with Avestan maha (Middle Persian mat, New Persian met), Greek 
fiedv, Latin temetum. Knowledge of grape-wine was conveyed to India 
from the West, as we see from the Periplus and Tamil poems alluding 
to the importation of Yavana (Greek) wines. 2 In the Raghuvanica 
(iv, 65), madhu doubtless refers to grape-wine; for King Raghu van- 
quished the Yavana, and his soldiers relieve their fatigue by enjoying 
madhu in the vine regions of the Yavana country. 

According to W. Ainslie, 3 the French at Pondicherry, in spite of the 
great heat of the Carnatic, are particularly successful in cultivating 
grapes; but no wine is made in India, nor is the fruit dried into raisins 
as in Europe and Persia. The Arabians and Persians, particularly the 
latter, though they are forbidden wine by the Koran, bestow much 
pains on the cultivation of the grape, and suppose that the different 
kinds possess distinguishing medicinal qualities. Wine is brought to 
India from Persia, where, according to Ta vernier (1605-89), three 
sorts are made: that of Yezd, being very delicate; the Ispahan produce, 
being not so good; and the Shiraz, being the best, rich, sweet, and 
generous, and being obtained from the small grapes called ki$mi$ t 
which are sent for sale to Hindustan when dried into raisins. 4 There 
are two brands of Shiraz wine, a red and a white, both of which are 
excellent, and find a ready market in India. Not less than four thou- 
sand tuns of Shiraz wine is said to be annually sent from Persia to 
different parts of the world. 5 The greatest quantity is produced in the 
district of Korbal, near the village of Bend Emir. 6 In regard to Assam, 

1 Compare above (p. 222) the wine of the Yue-ci. According to Boehtlingk, 
maireya is an intoxicating drink prepared from sugar and other substances. 

2 V. A. Smith, Early History of India, p. 444 (3d ed.). 

3 Materia Indica, Vol. I, p. 157. 

4 Compare above, p. 231. 

6 " Wines too, of every clime and hue, 

Around their liquid lustre threw; 
Amber Rosolli, — the bright dew 
From vineyards of the Green-Sea gushing; 
And Shiraz wine, that richly ran 
As if that jewel, large and rare, 
The ruby, for which Kublai-Khan 
Offer'd a city's wealth, was blushing 
Melted within the goblets there!" 

Thomas Moore, Lalla Rookh. 
6 Ainslee, I.e., p. 473. 

242 Sino-Iranica 

Ta vernier 1 states that there are quantities of vines and good grapes, 
but no wine, the grapes being merely dried to distil spirits from. Wild 
vine grows in upper Siam and on the Malay Peninsula, and is said to 
furnish a rather good wine. 8 

A wine-yielding plant of Central Asia is described in the Ku kin lu 
"6" 4* W" by Ts'ui Pao -^ 15 of the fourth century, as follows: "The 
tsiu-pei-t'en $f $£ M ("wine-cup creeper") has its habitat in the West- 
ern Regions (Si-yu). The creeper is as large as an arm; its leaves are 
like those of the ko 3§ (Pachyrhizus thunbergianus, a wild-growing 
creeper); flowers and fruits resemble those of the wu-t'un (Sterculia 
platanifolia) , and are hard; wine can be pressed out of them. The 
fruits are as large as a finger and in taste somewhat similar to the tou-k'ou 
a. H (Alpinia globosum) ; their fragrance is fine, and they help to digest 
wine. In order to secure wine, the natives get beneath the creepers, 
pluck the flowers, press the wine out, eat the fruit for digestion, and 
become intoxicated. The people of those countries esteem this wine, 
but it is not sent to China. Can K'ien obtained it when he left Ta-yuan 
(Fergana). This affair is contained in the Can K'ien I'u kwan H %fk If 
\ti §8 iS ('Memoirs of Can K'ien's Journey')." 4 This account is re- 
stricted to the Ku kin Zu, and is not confirmed by any other book. Li 
Si-cen's work is the only Pen ts'ao which has adopted this text in an 
abridged form. 6 Accordingly the plant itself has never been introduced 
into China; and this fact is sufficient to discard the possibility of an 
introduction by Can K'ien. If he had done so, the plant would have 
been disseminated over China and mentioned in the various early 
Pen ts'ao; it would have been traced and identified by our botanists. 
Possibly the plant spoken of is a wild vine, possibly another genus. 
The description, though by no means clear in detail, is too specific to 
be regarded as a mystification. 

The history of the grape-vine in China has a decidedly method- 
ological value. We know exactly the date of the introduction and 

x Travels in India, Vol. II, p. 282. 

9 Dilock Prinz von Siam, Landwirtschaft in Siam, p. 167. 

3 Ch. c, p. 2 b. The text has been adopted by the Sit po wu li (Ch. 5, p. 2 b) 
and in a much abbreviated form by the Yu yan tsa tsu (Ch. 18, p. 6 b). It is not in 
the Pen ts'ao kan mu, but in the Pen ts'ao kan mu Si i (Ch. 8, p. 27). 

* Hirth (Journal Am. Or. Soc, Vol. XXXVII, 1917, p. 91) states that this 
work is mentioned in the catalogue of the library of the Sui dynasty, but not in the 
later dynastic catalogues. We do not know when and by whom this alleged book 
was written; it may have been an historical romance. Surely it was not produced 
by Can K'ien himself. 

8 See also T'u Su tsi I'en, XX, Ch. 112, where no other text on the subject is 

The Grape-Vine 343 

the circumstances which accompanied this important event. We have 
likewise ascertained that the art of making grape-wine was not learned 
by the Chinese before a.d. 640. There are in China several species of 
wild vine which bear no relation to the imported cultivated species. 
Were we left without the records of the Chinese, a botanist of the 
type of Engler would correlate the cultivated with the wild forms and 
assure us that the Chinese are original and independent viticulturists. 
In fact, he has stated 1 that Vitis thunbergii, a wild vine occurring in 
Japan, Korea, and China, seems to have a share in the development of 
Japanese varieties of vine, and that Vitis filifolia of North China seems 
to have influenced Chinese and Japanese vines. Nothing of the kind 
can be inferred from Chinese records, or has ever been established by 
direct observation. The fact of the introduction of the cultivated grape 
into China is wholly unknown to Engler. The botanical notes appended 
by him to Hehn's history of the grape 1 have nothing whatever to do 
with the history of the cultivated species, but refer exclusively to wild 
forms. It is not botany, but historical research, that is able to solve the 
problems connected with the history of our cultivated plants. 

Dr. T. Tanaka of the Bureau of Plant Industry, U. S. Department 
of Agriculture, Washington, has been good enough to contribute the 
following notes on the history of the grape-vine in Japan: — 

"The early history of the cultivation of the grape-vine (Vitis 
vinifera) in Japan is very obscure. Most of the early Japanese medical 
and botanical works refer to bud flf %j (Chinese p'u-t'ao) as ebi, the 
name occurring in the Kojiki (compiled in a.d. 712, first printed in 
1644) as yebikadzura* which is identified by J. Matsumura 4 as Vitis 
vinifera. It seems quite incomprehensible that the grape-vine, which 
is now found only in cultivated form, should have occurred during the 
mythological period as early as 660 B.C. The Honzo-wamyd ^ 3£ 
bI %\ (compiled during the period 897-930, first printed 1796) mentions 
d-ebi-kadzura as vine-grape, distinguishing it from ordinary ebi-kadzura, 
but the former is no longer in common use in distinction from the latter. 
The ebi-dzuru which should correctly be termed inu-ebi (false ebi 
plant), as suggested by Ono Ranzan, 5 is widely applied in Japan for 
$£ J| (Chinese yin-yii), and is usually identified as Vitis thunbergii, 

1 Erlauterungen zu den Nutzpflanzen der gemassigten Zonen, p. 30. 

• Kulturpflanzen, pp. 85-91. 

• B. H. Chamberlain, Ko-ji-ki, p. xxxiv. 

• Botanical Magazine, Tokyo, Vol. VII, 1893, p. 139. 

• HontS komoku keimO, ed. 1847, Ch. 29, p. 3. 

244 Sino-Iranica 

but is an entirely different plant, with small, deeply-lobed leaves, 
copiously villose beneath. Ebi-kadzura is mentioned again in the 
Wamyo-ruijus'd ^0 ^ IS ^ ^ (compiled during the period 923-931, 
first edited in 161 7), which gives budo as the fruit of Ukwatsu or Vitis 
coignetiae 1 , as growing wild in northern Japan. 

"These three plants are apparently mixed up in early Japanese 
literature, as pointed out by Arai Kimiyosi. 2 Describing budo as a food 
plant, the Honlo Sokukan ^ 19 Jt HI 3 mentions that the fruit was not 
greatly appreciated in ancient times; for this reason no mention was 
made of it in the Imperial chronicles, nor has any appropriate Japanese 
term been coined to designate the vine-grape proper. 

"In the principal vine-grape district of Japan, Yamanasi-ken 
(previously called Kai Province), were found a few old records, an 
account of which is given in Viscount Y. Fukuba's excellent discourse 
on Pomology. 4 An article on the same subject was published by J. 
Dautremer. 6 This relates to a tradition regarding the accidental dis- 
covery by a villager, Amenomiya Kageyu (not two persons), of the vine- 
grape in 1 1 86 (Dautremer erroneously makes it 1195) at the mountain 
of Kamiiwasaki Ji ^ ^, not far from Kofu R? $f. Its cultivation must 
have followed soon afterward, for in 11 97 a few choice fruits were 
presented to the Sogun Yoritomo (1147-99). At the time of Takeda 
Harunobu (1521-73) a sword was presented to the Amenomiya family 
as a reward for excellent fruits which they presented to the Lord. 
Viscount Fukuba saw the original document relative to the official 
presentation of the sword, and bearing the date 1549. 6 The descendants 
of this historical grape-vine are still thriving in the same locality around 
the original grove, widely recognized among horticulturists as a true 
Vitis vinifera. According to a later publication of Fukuba, 7 there is 
but one variety of it. Several introductions of Vitis vinifera took place 
in the early Meiji period (beginning 1868) from Europe and America. 

"The following species of Vitis are mentioned in Umemura's work 
Inolokukwai-no-lokubutsu-li ffc 'ft <ft- iZ. fit tyli ffe 8 as being edible: 

1 Matsumura, Shokubutsu Mei-i, p. 380. 

2 Toga )ft #fl (completed in 1719), ed. 1906, p. 272. 
8 Ch. 4, p. 50 (ed. of 1698). 

4 Kwaju engei-ron ^ ^gj [§ §| f$f, privately published in 1892. 

6 Situation de la vigne dans l'empire du Japon, Transactions Asiatic Society of 
Japan, Vol. XIV, 1886, pp. 176-185. 

6 Fukuba, op. cit., pp. 461-462. 

7 Kwaju saibaijenSo y^ $$ ^ffc i § ■£ fj|, Vol. IV, 1896, pp. 1 19-120. 

8 Vol. 4, 1906. 

The Grape-Vine 245 

"Yama-budo (Vitis coignetiae): fruit eaten raw and used for wine; 
leaves substituted for tobacco. 

"Ebi-dzuru (V. thunbergii): fruit eaten raw, leaves cleaned and 
cooked; worm inside the cane baked and eaten by children as remedy 
for convulsions. 

" Sankaku-dzuru {V. flexuosa): fruit eaten raw. 

"Ama-dzuru (V. saccharifera) : fruit eaten raw; children are very 
fond of eating the leaves, as they contain sugar." 


3. Pistacia is a genus of trees or shrubs of the family Anacardiaceae, 
containing some six species, natives of Iran and western Asia, and also 
transplanted to the Mediterranean region. At least three species 
(Pistacia vera, P. terebinthus, and P. acuminata) are natives of Persia, 
and from ancient times have occupied a prominent place in the life of the 
Iranians. Pistachio-nuts are still exported in large quantities from 
Afghanistan to India, where they form a common article of food among 
the well-to-do classes. The species found in Afghanistan and Baluchis- 
tan do not cross the Indian frontier. 1 The pistachio (Pistacia vera) in 
particular is indigenous to ancient Sogdiana and Khorasan, 2 and still 
is a tree of great importance in Russian Turkistan. 3 

When Alexander crossed the mountains into Bactriana, the road 
was bare of vegetation save a few trees of the bushy terminthus or 
terebinthus. 4 On the basis of the information furnished by Alexander's 
scientific staff, the tree is mentioned by Theophrastus 5 as growing in 
the country of the Bactrians; the nuts resembling almonds in size 
and shape, but surpassing them in taste and sweetness, wherefore the 
people of the country use them in preference to almonds. Nicandrus 
of Colophon 6 (third century B.C.), who calls the fruit fiiaranov or ^lttolklov, 
a word derived from an Iranian language (see below), says that it grows 
in the valley of the Xoaspes in Susiana. Posidonius, Dioscorides, Pliny, 
and Galenus know it also in Syria. Vitellius introduced the tree into 
Italy; and Flaccus Pompeius, who served with him, introduced it at 
the same time into Spain. 7 

The youths of the Persians were taught to endure heat, cold, and 
rain; to cross torrents and to keep their armor and clothes dry; to 
pasture animals, to watch all night in the open air, and to subsist on 
wild fruit, as terebinths (Pistacia terebinthus), acorns, and wild pears. 8 

1 Watt, Dictionary of the Economic Products of India, Vol. VI, p. 268. 

2 Joret, Plantes dans l'antiquite\ Vol. II, pp. 47, 76. 

3 S. Korzinski, Vegetation of Turkistan (in Russian), pp. 20, 21. 

4 Strabo, XV. 11, 10. 

■ Hist, plant., IV. iv, 7. 
"Theriaka, 890. 

7 Pliny, xv, 22, §91. A. de Candolle (Origin of Cultivated Plants, p. 316) 
traces Pistacia vera only to Syria, without mentioning its occurrence in Persia. 

8 Strabo, XV. hi, 18. 


The Pistachio 247 

The Persians appeared to the ancients as terebinth-eaters, and this 
title seems to have developed into a sort of nickname: when Astyages, 
King of the Medians, seated on his throne, looked on the defeat of his 
men through the army of Cyrus, he exclaimed, "Woe, how brave are 
these terebinth-eating Persians!" 1 According to Polyaenus, 2 terebinth- 
oil was among the articles to be furnished daily for the table of the 
Persian kings. In the Bttndahisri, the pistachio-nut is mentioned to- 
gether with other fruits the inside of which is fit to eat, but not the 
outside. 3 "The fruits of the country are dates, pistachios, and apples 
of Paradise, with other of the like not found in our cold climate." 4 

Twan C'en-si It $ ^, in his Yu yah tsa tsu MiiS, written 
about a.d. 860 and containing a great amount of useful information 
on the plants of Persia and Fu-lin, has the following: — 

"The hazel-nut (Corylus heterophylla) of the Hu (Iranians), styled 
a-yiie H M , grows in the countries of the West. 6 According to the 
statement of the barbarians, a-yiie is identical with the hazel-nuts 
of the Hu. In the first year the tree bears hazel-nuts, in the second 
year it bears a-yiie." 6 

C'en Ts'an-k'i W W. II, who in the K'ai-yuan period (a.d. 713-741) 
wrote the Materia Medica Pen ts'ao H i ^ #• & 5ft, states that "the 
fruits of the plant a-yiie-hun M R W- are warm and acrid of flavor, 
non-poisonous, cure catarrh of the bowels, remove cold feeling, and 
make people stout and robust, that they grow in the western countries, 
the barbarians saying that they are identical with the hazel-nut of the 
Hu fifl $1 -?\ During the first year the tree bears hazel-nuts, in the 
second year it bears a-yiie-hun." 

Li Sun ^ ^J, in his Hat yao pen ts'ao M M ♦ ^ (second half of the 
eighth century), states, "According to the Nan lou ki ^f #H 12 by 
Su Piao fe $£, 7 the Nameless Tree (wu mih mu $& %\ /fc) grows in the 
mountainous valleys of Lin-nan (Kwan-tun) . Its fruits resemble in appear- 
ance the hazel-nut, and are styled Nameless Fruits (wu mih tse $& & 

1 Nicolaus of Damaskus (first century B.C.), cited by Hehn, Kulturpflanzen, 
p. 424. 

* Strategica, IV. 111, 32. 

8 These fruits are walnut, almond, pomegranate, coconut, filbert, and chestnut. 
See West, Pahlavi Texts, Vol. I, p. 103. 

4 Marco Polo, Yule's edition, Vol. I, p. 97. 

8 The editions of the Yu yan tsa tsu write © HI, "in the gardens of the West"; 
but the T'u Jm tsi I'en (section botany, Ch. 311) and Ci wu tnin H t'u k'ao, in repro- 
ducing this text, offer the reading ]§ S , which seems to me preferable. 

8 Yu yan tsa tsu jj|| ft, Ch. 10, p. 3 b (ed. of Tsin tat pi Su). 

7 This work is quoted in the Ts'i min yao Su, written by Kia Se-niu under the 
Hou Wei dynasty (a.d. 386-534). 

248 Sino-Iranica 

■?■). Persians 3£ $ft 1%. designate them a-yiie-hun fruits." 1 For the same 
period we have the testimony of the Arabic merchant Soleiman, who 
wrote in a.d. 851, to the effect that pistachios grow in China. 2 

As shown by the two forms, a-yiie of the Yu yan tsa tsu and a-yiie-hun 
of the Pen ts'ao H i and Hai yao pen ts'ao, the fuller form must repre- 
sent a compound consisting of the elements a-yiie and hun. In order to 
understand the transcription a-yiie, consideration of the following facts 
is necessary. 

The Old-Iranian word for the walnut has not been handed down to 
us, but there is good evidence to prompt the conclusion that it must 
have been of the type *ag5za or *ahgOza. On the one hand, we have 
Armenian engoiz, Ossetic angozii or anguz, and Hebrew egdz; 3 on the 
other hand, we meet in Yidgha, a Hindu-Kush language, the form 
ogiizo, as compared with New Persian koz and g5z. A The signification 
of this word is "nut" in general, and "walnut" in particular. Further, 
there is in Sanskrit the Iranian loan-word akhdta, aksofa, or aksoda, 
which must have been borrowed at an early date, as, in the last-named 
form, the word occurs twice in the Bower Manuscript. 5 It has survived 
in Hindustani as axrot or dkrot. The actual existence of an East- 
Iranian form with the ancient initial a- is guaranteed by the Chinese 
transcription a-yiie; for a-yiie M M answers to an ancient *a-hwie5 
(hw'e5) or *a-gwie5, a-gwii5; 6 and this, in my opinion, is intended to 
represent the Iranian word for "nut" with initial a-, mentioned above; 
that is, *ahgwlz, angwOz, agOz. 

Chinese hun W answers to an ancient *7wun or wun. In regard 
to this Iranian word, the following information may be helpful. E. 

1 If it is correct that the transcription a-yiie-hun was already contained in the 
Nan iou ki (which it is impossible to prove, as we do not possess the text of this 
work), the transcription must have been based on an original prototype of early 
Sasanian times or on an early Middle-Persian form. This, in fact, is confirmed by 
the very character of the Sino- Iranian word, which has preserved the initial a-, 
while this one became lost in New Persian. It may hence be inferred that Li Sun's 
information is correct, and that the transcription a-yiie-hun may really have been 
contained in the Nan cou ki, and would accordingly be pre-T'an. 

2 M. Reinaud, Relation des voyages faits par les Arabes et les Persans dans 
l'lnde et a la Chine, Vol. I, p. 22. 

3 Whether Georgian nigozi and the local name Nf-youf a of Ptolemy (W. 
Tomaschek, Pamirdialekte, Sitzber. Wiener Akad., 1880, p. 790) belong here, I do 
not feel certain. Cf. HObschmann, Armenische Grammatik, p. 393. 

4 In regard to the elision of initial a in New Persian, see Hubschmann, Persische 
Studien, p. 120. 

8 Hoernle's edition, pp. 32, 90, 121. 

8 Regarding the phonetic value of ^ , see the detailed study of Pelliot (Bull, 
de I'Ecole frangaise, Vol. V, p. 443) and the writer's Language of the Yue-chi or 

The Pistachio 249 

Kaempfer 1 speaks of Terebinthus or Pistacea syhestris in Persia thus: 
"Ea Pistaceae hortensi, quam Theophrastus Therebinthum Indicam 
vocat, turn magnitudine, turn totius ac partium figura persimilis est, 
nisi quod flosculos ferat fragrantiores, nuces vero praeparvas, insipidas; 
unde a descriptione botanica abstinemus. Copiosa crescit in recessibus 
montium brumalis genii, petrosis ac desertis, circa Schamachiam Mediae, 
Schirasum Persidis, in Luristano et Larensi territoriis. Mihi nullibi 
conspecta est copiosior quam in petroso monte circa Majin, pagum 
celebrem, una diaeta dissitum Sjiras6: in quo mihi duplicis varietatis 
indicarunt arborem; unam vulgariorem, quae generis sui retineat 
appellationem Diracht [diraxt, 'tree'] Ben seu Wen; alteram rariorem, 
in specie Kasudaan [kasu-dan], vel, ut rustici pronunciant, Kasudln 
dictam, quae a priori fructuum rubedine differat." Roediger and Pott 2 
have added to this ben or wen a Middle-Persian form ven ("wild pista- 
chio"). In the Persian Dictionary edited by Steingass (p. 200) this 
word is given as ban or wan (also banak), with the translation "Persian 
turpentine seed." 3 Vullers 4 writes it ban. Schlimmer 5 transcribes 
this word beneh. He identifies the tree with Pistacia acuminata and 
observes, "C'est l'arbre qui fournit en Perse un produit assez semblable 
a la tr^mentine, mais plut6t mou que liquide, vu qu'on l'obtient par 
des d£coupures, dont le produit se rassemble durant les grandes chaleurs 
dans un creux fait en terre glaise au pied de l'arbre, de facon a ce que la 
matiere s£cr£t£e perd une grande partie de son huile essentielle avant 
d'etre enlev£e. Le mSme produit, obtenu a Kerman dans un outre, 
fixe" a l'arbre et enleve" aussit6t plein, £tait a peu pres aussi liquide que 
la ter^benthine de Venise. ... La Pistacia acuminata est sauvage au 
Kordesthan persan et, d'apr^s Buhse, aussi a Reshm, Damghan et 
Dereghum (province de Yezd) ; Haussknecht la vit aussi a Kuh Kiluye 
et dans le Luristan." 

The same word we meet also in Kurd dariben, dar-i-ben ("the tree 
ben"), and in all probability in Greek Tepkflivdos, older forms repfuvdos 
and rpkfudos. 6 Finally Watt 7 gives a Balu£l word ban, wan, wana, gwa, 

1 Amoenitatum exoticarum fasciculi V, p. 413 (Lemgoviae, 1712). 

2 Zeitschr. Kunde d. Morgenl., Vol. V, 1844, p. 64. 

s This notion is also expressed by ban&sib (cf. binast, "turpentine"). 
* Lexicon persico-latinum, Vol. I, p. 184. 

6 Terminologie, p. 465. 

*The Greek ending, therefore, is -Bo%, not -vOm, as stated by Schrader (in 
Hehn, Kulturpflanzen, 8th ed., p. 221); n adheres to the stem: lere-bin-dos. 

7 Commercial Products of India, p. 902; and Dictionary of the Economic 
Products of India, Vol. VI, p. 271. 

t$o Sino-Iranica 

gwan, gwana, for Pistacia mutica (or P. terebinthus, var. mutica); this 
form comes nearest to the Chinese transcription. 

While a compound *agoz-van(vun), that is, "nut of pistachio," as 
far as I know, has not yet been traced in Iranian directly, its existence 
follows from the Chinese record of the term. An analogy to this com- 
pound is presented by Kurd kizvan, kezvan, kazu-van, kasu-van ("pista- 
chio" or "terebinthus-tree"). 1 

The Honzo kdmoku keimo (Ch. 25, fol. 24), written by Ono Ranzan 
/J» I? W tfj, first published in 1804, revised in 1847 by IguSi Bosi # 
n !|? £,, his grandson, mentions the same plant ffl ft 'W- -J", which 
reads in Japanese agetsu-konU. He gives also in Kana the names 
fusudasiu or fusudasu. 2 He states, "The plant is not known in Japan 
to grow wild. It used to come from foreign countries, but not so at 
present. A book called Zdkyohi furoku % ffc 2& Pft $& mentions this 
plant, stating that agetsu-konH is the fruit of the tree £'a mu M Js 
(in Japanese sakuboku)." 3 

1 A. Jaba, Dictionnaire kurde-francais, p. 333. Cf. above the kasu-dan of 

* These terms are also given by the eminent Japanese botanist Matsumura 
in his Shokubutsu mei-i (No. 2386), accompanied by the identification Pistacia 

* This tradition is indeed traceable to an ancient Chinese record, which will be 
found in the Cen lei pen ts'ao of 1108 (Ch. 12, p. 55, ed. of 1583). Here the question 
is of the bark of the san or Va tree flffl ;fv fj£, mentioned as early as the sixth century 
in the Kwan li Hf ^ of Kwo Yi-kun as growing in wild country of Kwah-nan 
^ f$3 (the present province of Kwah-tun and part of K wan-si), and described in a 
commentary of the Er ya as resembling the mulberry-tree. This, of course, is a wild 
tree indigenous to a certain region of southern China, but, as far as I know, not yet 
identified, presumably as the ancient name is now obsolete. The Nan lou ki by 
Su Piao (see above) says that the fruits of this tree are styled wu min tse $H ^ J- 
(" nameless fruits"); hence the conclusion is offered by T'an Sen-wei, author of the 
Cen lei pen ts'ao, that this is the tree termed a-yile-hun by the Persians (that is, a cul- 
tivated Pistacia). This inference is obviously erroneous, as the latter was introduced 
from Persia into China either under the T'ang or a few centuries earlier, while the 
san or Va tree pre-existed spontaneously in the Chinese flora. The only basis for this 
hazardous identification is given by the attribute "nameless." A solution of this 
problem is possible if we remember the fact that there is a wild Pistacia, Pistacia 
chinensis, indigenous to China, and if we identify with it the tree san or Va; then it 
is conceivable that the wild and the imported, cultivated species were correlated 
and combined under the same popular term wu min. Matsumura (op. cit., No. 
2382) calls P. chinensis in Japanese orenju, adding the characters ^ f^. The word 
lien refers in China to Melia azedarach. The modern Chinese equivalent for P. 
chinensis is not known to me. The peculiar beauty of this tree, and the great age to 
which it lives, have attracted the attention of the indefatigable workers of our 
Department of Agriculture, who have already distributed thousands of young trees to 
parks throughout the country (see Yearbook of the U. S. Department of Agriculture 
1916, p. 140, Washington, 191 7). In the English and Chinese Standard Dictionary, 
the word "pistachio" is rendered by fei HI, which, however, denotes a quite dif- 

The Pistachio 251 

G. A. Stuart 1 has identified a-yiie hun-tse 1 with Pistacia vera, and 
this is confirmed by Matsumura. 

The Japanese name fusudasiu or fusudasu is doubtless connected 
with Persian pista, from Old Iranian *pistaka, Middle Persian *pistak, 8 
from which is derived Greek /SicrdKiov, ^ittclkiov, TnaraKiov or ^iotcucioj', 
Latin psittacium, and our pistacia or pistachio. It is not known to me, 
however, to what date the Japanese word goes back, or through what 
channels it was received. In all likelihood it is of modern origin, the 
introduction into Japan being due to Europeans. 

In Chinese literature, the Persian word appears in the Geography 
of the Ming Dynasty, 4 in the transcription [ki-] pi-se-tan [f5] $$> J8 fit, 
stated to be a product of Samarkand, the leaves of the tree resembling 
those of the San l x a tfj & {Camellia oleifera), and its fruit that of the 
yin hin IB ^ {Salisburia adiantijolia) . 

The Persian word, further, occurs in the new edition of the Kwan yii 
ki, entitled Tseh tin kwan yii ki *£ IT J^ H IE. The original, the Kwan 
yii ki, was written by Lu Yin-yan |££ M ^, 5 and published during the 
Wan-li period in 1600. The revised and enlarged edition was prepared 
by Ts'ai Fan-pin W: fS Jfi (hao Kiu-hia % it) in 1686; a reprint of 
this text was issued in 1744 by the publishing-house Se-mei fan 0H^. 
Both this edition and the original are before me. The latter 6 mentions 
only three products under the heading "Samarkand"; namely, coral, 
amber, and ornamented cloth {hwa hii pujk$& ^). The new edition, 
however, has fifteen additional items, the first of these being [ki-] 
pi-se-t'an, written as above, 7 stated to be a tree growing in the region 
of Samarkand. "The leaves of the tree," it is said, "resemble those 
of the San Pa {Cornelia oleifera) ; the fruits have the appearance of the 
nut-like seeds of the yin hin {Salisburia adiantijolia), but are smaller." 
The word pi-se-Van doubtless represents the transcription of Persian 

ferent plant, — Torreya nucifera. A revival on the part of the Chinese, of the good, 
old terms of their own language, would be very desirable, not only in this case, but 
likewise in many others. 

1 Chinese Materia Medica, p. 334. 

s Wrongly transcribed by him o-yiieh-chiin-tzu. 

3 These reconstructions logically result from the phonetic history of Iranian, 
and are necessitated by the existence of the Greek loan-word. Cf., further, Byzantine 
pustux and fustox, Comanian pistac, and the forms given below (p. 252). Persian 
pista is identified with Pistacia vera by Schlimmer (Terminologie, p. 465). 

4 Ta Min i t'un U, Ch. 89, p. 23. 

5 Wylie, Notes on Chinese Literature, p. 59. 

• Ch. 24, p. 6 b. 

7 The addition of ki surely rests on an error (Schott also reads pi-se-t'an, which 
he presumably found in his text; see the following note). 

252 Sino-Iranica 

pistan ("a place abounding with pistachio-nuts"). 1 Again, the Persian 
word in the transcription pi-se-ta >Sb A %£ appears in the Pen ts'ao 
kan mu H i 2 by Cao Hio-min, who states that the habitat of the plant 
is in the land of the Mohammedans, and refers to the work Yin San 
leh yao 3 of 133 1, ascribed by him to Hu-pi-lie %& 4b Wt\ that is, the 
Emperor Kubilai of the Yuan dynasty. We know, however, that this 
book was written in 133 1 by Ho Se-hwi. 4 Not having access to this, 
I am unable to state whether it contains a reference to pi-se-ta, nor do 
I know whether the text of Cao Hio-min, as printed in the second 
edition of 1765, was thus contained in the first edition of his work, which 
was published in 1650. It would not be impossible that the tran- 
scription pi-se-ta, accurately corresponding to Persian pista, was 
made in the Mongol period; for it bears the ear-marks of the Yuan style 
of transcription. 

The Persian word pista (also pasta) has been widely disseminated: 
we find it in Kurd fystiq, Armenian fesdux and jstoul, Arabic fistaq or 
fustaq, Osmanli fistiq, 6 and Russian fistaZka. 

In the Yuan period the Chinese also made the acquaintance of 
mastic, the resinous product of Pistacia lentiscus* It is mentioned in 
the Yin ia« leh yao, written in 133 1, under its Arabic name mastaki, 
in the transcription $1 JS %£ l=f ma-se-ta-ki. 7 Li Si-cen knew only the 
medical properties of the product, but confessed his ignorance regarding 
the nature of the plant; hence he placed his notice of it as an appendix 
to cummin {M-lo). The Wu tsa tolll, written in 1610, says that 
mastaki is produced in Turkistan and resembles the tsiao W> (Zanth- 
oxylum, the fruit yielding a pepper-like condiment) ; its odor is very 
strong; it takes the place there of a condiment like pepper, and is 
beneficial to digestion. 8 The Persian word for "mastic" is kundurak 
(from kundur, "incense"), besides the Arabic loan-word mastaki or 

1 As already recognized by W. Schott (Topographie der Producte des chinesi- 
schen Reiches, Abh. Berl.Akad., 1842, p. 371), who made use only of the new edition. 

2 Ch. 8, p. 19; ed. of 1765 (see above, p. 229). 
8 Cf. above, p. 236. 

4 Bretschneider, Bot. Sin., pt. 1, p. 213. 

6 Hence Pegoletti's fistuchi (Yule, Cathay, new ed. by Cordier, Vol. Ill, 
p. 167). 

6 Greek axivos (Herodotus, iv, 177). 

7 The Arabic word itself is derived from Greek naorlxn (from paor&feu', "to 
chew"), because the resin was used as a masticatory. Hence also Armenian tnaz- 
tak'e. Spanish almdciga is derived from the Arabic, as indicated by the Arabic 
article al, while the Spanish form m&sticis is based on Latin mastix. 

8 Quoted in the Pen ts'ao kan mu H t, Ch. 6, p. 12 b. The digestive property 
is already emphasized by Dioscorides (1, 90). 

The Pistachio 253 

mastaki. 1 The Persianized form is masdax; in Kurd it is mstekki. "On 
these mountains the Mastich Tree brings forth plenty of that gum, of 
which the country people make good profit. ... As for the Mastick 
Trees, they bore red berries, and if wounded would spew out the liquid 
resin from the branches; they are not very tall, of the bigness of our 
Bully Trees: Whether they bring forth a cod or not, this season would 
not inform me, nor can I say it agrees in all respects with the Lentisk 
Tree of Clusius." 2 The resin (mastic) occurs in small, irregular, yellowish 
tears, brittle, and of a vitreous fracture, but soft and ductile when 
chewed. It is used as a masticatory by people of high rank in India to 
preserve the teeth and sweeten the breath, and also in the preparation 
of a perfume. 8 It is still known in India as the "gum mastic of Rum." 4 
The case of the pistachio (and there are several others) is interesting 
in showing that the Chinese closely followed the development of Iranian 
speech, and in course of time replaced the Middle-Persian terms by the 
corresponding New-Persian words. 

1 Achundow, Abu Mansur, pp. 137, 267. 

* John Fryer, New Account of East India and Persia, Vol. II, p. 202 (Hakluyt 
Soc, 1912). 

' Watt, Commercial Products of India, p. 902. 

4 D. C. Phillott, Journal As. Soc. Bengal, Vol. VI, 1910, p. 81. 


4. The Buddhist dictionary Fan yi min yi tsi ffli M %\ Wi $&, 
compiled by Fa Yun ££ S, 1 contains a Chinese-Sanskrit name for the 
walnut (hu t'ao t$ $ls, Juglans regia) in the transcription po-lo-H 
IS "SI ($, which, as far as I know, has not yet been identified with its 
Sanskrit equivalent. 2 According to the laws established for the Buddhist 
transcriptions, this formation is to be restored to Sanskrit pdrasi, 
which I regard as the feminine form of the adjective pdrasa, meaning 
"Persian" (derived from Parsa, "Persia"). The walnut, accordingly, 
as expressed by this term, was regarded in India as a tree or fruit sus- 
pected of Persian provenience. The designation pdrasi for the walnut 
is not recorded in Boehtlingk's Sanskrit Dictionary, which, by the way, 
contains many other lacunes. The common Sanskrit word for "walnut " 
is dkhofa, aksdta, aksosa, 3 which for a long time has been regarded as 
a loan-word received from Iranian. 4 

Pliny has invoked the Greek names bestowed on this fruit as testi- 
mony for the fact that it was originally introduced from Persia, the 

1 Ch. 24, p. 27 (edition of Nanking). — Bunyiu Nanjio (Catalogue of the 
Buddhist Tripitaka, No. 1640) sets the date of the work at 1151. Wylie (Notes on 
Chinese Literature, p. 210) and Bretschneider (Bot. Sin., pt. 1, p. 94) say that it 
was completed in 1143. According to S. Julien (Methode, p. 13), it was compiled 
from 1 143 to 1 157. 

1 Bretschneider (Study and Value of Chinese Botanical Works, Chinese 
Recorder, Vol. Ill, 1871, p. 222) has given the name after the Pen ts'aokan mu, but 
has left it without explanation. 

* The last-named form occurs twice in the Bower Manuscript (Hoernle's 
edition, pp. 32, 90, 121). In Hindustani we have axrot or akrot. 

4 F. Spiegel, Arische Periode, p. 40. The fact that the ancient Iranian name for 
the walnut is still unknown does not allow us to explain the Sanskrit word satisfac- 
torily. Its relation to Hebrew egoz, and Persian koz, goz (see below), is perspicuous. 
Among the Hindu- Kush languages, we meet in Yidgha the word oghuzoh (J. Biddulph, 
Tribes of the Hindoo Koosh, Appendices, p. clxvii), which appears as a missing 
link between Sanskrit on the one hand and the Semitic-Armenian forms on the other 
hand: hence we may conjecture that the ancient Iranian word was something like 
*agoza, afigoza; and this supposition is fully confirmed by the Chinese transcription 
a-yiie (above, p. 248). Large walnuts of India are mentioned by the traveller C'an 
Te toward the middle of the thirteenth century (Bretschneider, Mediaeval 
Researches, Vol. I, p. 146). The walnuts of the province of Kusistan in Persia, which 
are much esteemed, are sent in great quantities to India (W. Ainslie, Materia 
Indica, Vol. I, p. 464). 


The Walnut 255 

best kinds being styled in Greek Persicum and basilicon, 1 and these being 
the actual names by which they first became known in Italy. 1 Pliny 
himself employs the name nuces iuglandes. Although Juglans regia is 
indigenous to the Mediterranean region, the Greeks seem to have 
received better varieties from anterior Asia, hence Greek names like 
K&pva Trepauia or napva <nvo)Tiica. 

In fact, Juglans regia grows spontaneously in northern Persia and 
in Baluchistan; it has been found in the valleys of the Pskem and 
Ablatun at altitudes varying from 1000 to 1500 m. Another species 
(Juglans pterocarpa, "Juglans with winged fruits") is met in the prov- 
inces of Ghilan and Mazanderan and in the vicinity of Astrabad. 4 
A. Engler 8 states that the walnut occurs wild also in eastern Afghanis- 
tan at altitudes of from 2200 to 2800 m. Ibn Haukal extols the walnuts 
of Arrajan, MuqaddasI those of Kirman, and Istaxri those of the 
province of Jlruft. 6 

In Fergana, Russian Turkistan, the walnut is cultivated in gardens; 
but the nuts offered for sale are usually derived from wild-growing trees 
which form complete forests in the mountains. 7 According to A. Stein, 8 
walnuts abound at Khotan. The same explorer found them at Yul-arik 
and neighboring villages. 6 

1 That is, "Persian nut" and "nut of the king," respectively, the king being 
the Basileus of Persia. These two designations are also given by Dioscorides (1, 178). 

2 Et has e Perside regibus translatas indicio sunt Graeca nomina: optimum 
quippe genus earum Persicum atque basilicon vocant, et haec fuere prima nomina 
{Nat. hist., xv, 22, § 87). 

• J. Hoops, Waldbaume und Kulturpflanzen, p. 553. The Romans transplanted 
the walnut into Gallia and Germania during the first centuries of our era. Numerous 
walnuts have been brought to light from the wells of the Saalburg, testifying to 
the favor in which they were held by the Romans. The cultivation of the tree is 
commended in Charles the Great's Capitulate de villis and Garden Inventories. 
Its planting in Gaul is shown by the late Latin term nux gallica, Old French nois 
gauge, which survives in our "walnut" (German walnuss, Danish valnod, Old Norse 
valhnot, Anglo-Saxon wealh-hnutu) ; walk, wal, was the Germanic designation of the 
Celts (derived from the Celtic tribe Volcae), subsequently transferred to the Romanic 
peoples of France and Italy. 

• C. Joret, Plantes dans I'antiquitl, Vol. II, p. 44. Joret (p. 92) states that the 
Persians cultivated nut-trees and consumed the nuts, both fresh and dried. The 
walnut is twice mentioned in the BQndahisn among the fruits serving as food, and 
among fruits the inside of which is fit to eat, but not the outside (West, Pahlavi 
Texts, Vol. I, pp. 101, 103; cf. also p. 275). 

8 Erlauterungen zu den Nutzpflanzen der gemassigten Zonen, p. 22. 

• P. Schwarz, Iran im Mittelalter, pp. 114, 218, 241. 

7 S. Korzinski, Sketches of the Flora of Turkistan, in Russian {Memoirs Imp. 
Russ. Ac, 8th ser., Vol. IV, No. 4, pp. 39, 53). 

8 Ancient Khotan, Vol. I, p. 131. 

• Ruins of Desert Cathay, Vol. I, p. 152. 

256 Sino-Iranica 

The New-Persian name for the walnut is kdz and goz. 1 According 
to Hubschmann, this word comes from Armenian. 2 The Armenian word 
is Ingoiz; in the same category belongs Hebrew egoz, 3 Ossetic angoza, 
Yidghal oyuza, Kurd egviz, Gruzinian nigozi. 4 The Persian word we 
meet as a loan in Turkish koz and xoz. h 

The earliest designation in Chinese for the cultivated walnut is hu 
fao tft $fc (" peach of the Hu": Hu being a general term for peoples of 
Central Asia, particularly Iranians). As is set forth in the Introduction, 
the term hu ijs prefixed to a large number of names of cultivated plants 
introduced from abroad. The later substitution hu or ho fao %% $ls 
signifies "peach containing a kernel," or "seed-peach," so called because, 
while resembling a peach when in the husk, only the kernel is eaten. 6 
In view of the wide dissemination of the Persian word, the question 
might be raised whether it would not be justifiable to recognize it also 
in the Chinese term hu fao f$ t&, although, of course, in the first line it 
means "peach of the Hu (Iranians)." There are a number of cases 
on record where Chinese designations of foreign products may simulta- 
neously convey a meaning and represent phonetic transcriptions. 
When we consider that the word hu ^B was formerly possessed of an 
initial guttural sonant, being sounded *gu (7U) or *go, 7 the possibility 
that this word might have been chosen in imitation of, or with especial 
regard to, an Iranian form of the type goz, cannot be denied: the two- 
fold thought that this was the "peach styled go" and the "peach of the 
Go or Hu peoples" may have been present simultaneously in the minds 
of those who formed the novel term; but this is merely an hypothesis, 
which cannot actually be proved, and to which no great importance is 
to be attached. 

1 Arabic joz; Middle Persian joz, joj. Kurd gwiz (guwiz), from govz, gdz (Socin, 
Grundr. iran. Phil., Vol. I, pt. 2, p. 268). Sariqoll ghauz (Shaw, Journal As. Soc. 
Bengal, 1876, p. 267). Pu§tu ughz, waghz. Another Persian designation for " walnut " 
is girdu or girdgdn. \ 

2 Grundr. iran. Phil., Vol. I, pt. 2, p. 8; Armen. Gram., p. 393. 
8 Canticle vi, 10. Cf. Syriac gauze. 

* W. Miller, Sprache der Osseten, p. 10; Hubschmann, Arm. Gram., p. 393. 

5 Radloff, Wdrterbuch der Turk-Dialecte, Vol. II, col. 628, 1710. In Osmanli 


8 The term ho t'ao is of recent date. It occurs neither under the T'ang nor 
under the Sung. It is employed in the Kwo su ^ 0ft, a work on garden-fruits by 
Wan Si-mou 3* "tft S£. who died in 1591, and in the Pen ts'ao kan mu. The latter 
remarks that the word ho ^ is sounded in the north like hu ffl , and that the sub- 
stitution thus took place, citing a work Min wu li & tyft jj> as the first to apply 
this term. 

7 Compare Japanese go-ma ft^ jjjjc and go-fun ffl ffi. 

The Walnut 257 

There is a tradition to the effect that the walnut was introduced 
into China by General Can K'ien. 1 This attribution of the walnut to 
Can K'ien, however, is a purely retrospective thought, which is not 
contained in the contemporaneous documents of the Han Annals. There 
are, in fact, as we have seen, only two cultivated plants which can 
directly be credited to the mission of Can K'ien to the west, — the 
grape and the alfalfa. All others are ascribed to him in subsequent 
books. Bretschneider, in his long enumeration of Can-K'ien plants, 2 
has been somewhat uncritical in adopting the statements of such a 
recent work as the Pen ts'ao kah mu without even taking pains to ex- 
amine the sources there referred to. This subject requires a renewed 
critical investigation for each particular plant. As regards the walnut, 
Bretschneider was exposed to singular errors, which should be rectified, 
as they have passed into and still prominently figure in classical botani- 
cal and historical books of our time. According to Bretschneider, the 
walnut was brought from K'iang-hu ifi SH, and "K'iang" was at the 
time of the Han dynasty the name for Tibet. There is, of course, no 
such geographical name as "K'iah-hu"; but we have here the two 
ethnical terms, "K'ian" and "Hu," joined into a compound. More- 
over, the K'iah (anciently *Gian) of the Han period, while they may 
be regarded as the forefathers of the subsequent Tibetan tribes, did 
not inhabit the country which we now designate as Tibet; and the term 
"Hu" as a rule does not include Tibetans. What is said in this respect 
in the Pen ts'ao kah mu 3 is vague enough : it is a single sentence culled 
from the T % u kin pen ts x ao H M. ^ W of Su Sun M $& (latter part of 
the eleventh century) of the Sung period, which reads, "The original 
habitat of this fruit was in the countries of the K'ian and the Hu" 
(jft. l& ^ ffi j& fifl). Any conclusion like an introduction of the walnut 
from "Tibet" cannot be based on this statement. 

Bretschneider's first victim was the father of the science of historical 
and geographical botany, A. de Candolle, 4 who stated, referring to 
him as his authority, "Chinese authors say that the walnut was 
introduced among them from Tibet, under the Han dynasty, by Chang- 

1 The first to reveal this tradition from the Pen ts'ao kan mu was W. Schott 
(Abh. Berl. Akad., 1842, p. 270). 

1 Chinese Recorder, 1871, pp. 221-223; and Bot. Sin., pt. I, p. 25. Likewise 
Hirth, T'oung Pao, Vol. VI, 1895, p. 439. Also Giles (Biographical Dictionary, p. 12) 
connects the walnut with Can K'ien. 

8 Ch. 30, p. 16. 

4 Origin of Cultivated Plants, p. 427. 

258 Sino-Iranica 

kien, about the year 140-150 B.C." 1 In Hehn's "Kulturpnanzen" 2 
we still read in a postscript from the hand of the botanist A. Engler, 
"Whether the walnut occurs wild in North China may be doubted, as 
according to Bretschneider it is said to have been imported there from 
Tibet." As will be seen below, a wild-growing species of Juglans is 
indeed indigenous to North China. As to the alleged feat of Can K'ien, 
the above-mentioned Su Sufi, who lived during the Sung period in the 
latter part of the eleventh century, represents the source of this purely 
traditional opinion recorded by Bretschneider. Su Sun, after the above 
statement, continues, "At the time of the Han, when Can K'ien was 
sent on his mission into the Western Regions, he first obtained the 
seeds of this fruit, which was then planted in Ts'in (Kan-su) ; at a later 
date it gradually spread to the eastern parts of our country; hence it 
was named hu Vao." z Su Sufi's information is principally based on the 
Pen ts'ao of the Kia-yu period (1056-64) M tf& W t£ 4* W-; this work 
was preceded by the Pen ts'ao of the K'ai-pao period (968-976) PD St 
# ^; and in the latter we meet the assertion that Can K'ien should 
have brought the walnut along from the Western Regions, but cautiously 
preceded by an on dit (z*). 4 The oldest text to which I am able to trace 
this tradition is the Po wu U fil tyi ;S of Can Hwa §6 W (a.d. 23 2-3 00). 6 
The spurious character of this work is well known. The passage, at any 
rate, existed, and was accepted in the Sung period, for it is reproduced 
in the T'ai pHn yii Ian} We even find it quoted in the Buddhist dic- 
tionary Yi tsHe kin yin t — ty) H£ if i§, 7 compiled by Yuan Yin jt JS8 
about a.d. 649, so that this tradition must have been credited in the 

1 Besides Bretschneider's article in the Chinese Recorder, de Candolle refers to 
a letter of his of Aug. 23, 1881, which shows that Bretschneider had not changed 
his view during that decade. Needless to add, that Can K'ien never was in Tibet, 
and that Tibet as a political unit did not exist in his time. Two distinct traditions 
are welded together in Bretschneider's statement. 

' Eighth edition (191 1), p. 400. \ 

3 Cert lei pen ts'ao, Ch. 23, p. 45 (edition of 152 1). G. A. Stuart (Chinese 
Materia Medica, p. 223) regards the "Tangut country about the Kukunor" as the 
locality of the tree pointed out in the Pen ts'ao. 

* The text of the K'ai-pao pen ts'ao is not reproduced in the Pen ts'ao kan mu> 
but will be found in the Ci wu min Si t'u k'ao, Ch. 17, p. 33. T'ah Sen-wei ^ ty| ftfc> 
in his Cen lei pen ts'ao (Ch. 23, p. 44 b), has reproduced the same text in his own 

6 §g H ^ 6 £ m. Jb (or jg) # M £ U (Ch. 6, p. 4, of the Wu-c'an 

•Ch. 971, p. 8. 

7 Ch. 6, p. 8 b (ed. of Nanking). In this text the pomegranate and grape are 
added to the walnut. In the same form, the text of the Po wu li is cited in the modern 
editions of the Ts'i min yao Su (Ch. 10, p. 4). 

The Walnut 259 

beginning of the T'ang dynasty. It is not impossible, however, that 
this text was actually written by Can Hwa himself, or at least that the 
tradition underlying it was formed during the fourth century; for, as 
will be seen, it is at that time that the walnut is first placed on record. 
Surely this legend is not older than that period, and this means that 
it sprang into existence five centuries after Can K'ien's lifetime. It 
should be called to mind that the Po wu li entertains rather fantastic 
notions of this hero, and permits him to cross the Western Sea and even 
to reach Ta Ts'in. 1 It is, moreover, the Po wu U which also credits to 
Can K'ien the introduction of the pomegranate and of ta or hu swan 
sfc (tfl ) i^ or hu $1 (Allium scorodoprasum) . 2 Neither is this tradition 
contained in the texts of the Han period. The notion that Can K'ien 
really introduced the walnut in the second century B.C. must be posi- 
tively rejected as being merely based on a retrospective and unauthentic 
account. 3 

The question now arises, Is there any truth in Su Sun's allegation 
that the walnut was originally produced in the country of the K'iah? 
Or, in other words, are we entitled to assume the co-existence of two 
Chinese traditions, — first, that the walnut was introduced into China 
from the regions of the Hu (Iranians) ; and, second, that another intro- 
duction took place from the land of the K'ian, the forefathers of the 
Tibetans?* There is indeed an ancient text of the Tsin period from the 
first part of the fourth century, one of the earliest datable references 
to the walnut, in which its origin from the K'ian is formally admitted. 
This text is preserved in the T'ai pHh yii Ian as follows: — 

"The mother of Liu T'ao £>J i§, 5 in her reply to the letter of Yu 
§1 , princess of the country of Wu ^| S3 , said, ' In the period Hien-ho 
J& IP (a.d. 326-335, of the Tsin dynasty) I escaped from the rebellion 

»Ch. 1, p. 3 b. 
1 See below, p. 302. 

* The Can-K'ien legend is also known in Korea (Korea Review, Vol. II, 1902, 
P- 393). 

4 The term k'iah t'ao 7^ $|j for the walnut is given, for instance, in the Hwa 
ki* ~%C H • "Mirror of Flowers" (Ch. 3, p. 49), written by C'en Hao-tse $£ /f| 
J* in 1688. He gives as synonyme also wan swi tseffi Wl "?■ ("fruits of ten thousand 
years"). The term k'ian t'ao is cited also in the P'ex wen lai kwan k'iin fan p'u 
(Ch. 58, p. 24; regarding this work cf. Bretschneider, Bot. Sin., pt. 1, p. 70), and in 
the P'an San li *}SL tfj *S (Ch.15, p. 2 b; published in 1755 by order of K'ien-lun). 

5 The T'u Su tsi I'en and Kwan k'iin fan p'u (Ch. 58, p. 25) write this name Niu 
jjaft. The Ko li kin yuan (Ch. 76, p. 5), which ascribes this text to the Tsin Su, gives 
it as ££. The Tan Sun pai k'un leu Vie M 5|c & ?L /\ OH (Ch. 99, p. 12) has, "The 
mother of Liu T'ao of the Tsin dynasty said, in reply to a state document, 'walnuts 
were originally grown in the country of the Western K'iah.'" 

260 Sino-Iranica 

of Su Tsun 1£ $t l into the Lin-nan mountains $8> ^c \U. The country 
of Wu sent a messenger with provisions, stating in the accompanying 
letter: 'These fruits are walnuts $1 $li and fei-San M IS. 2 The latter 
come from southern China. The walnuts were originally grown abroad 
among the Western K'iah (^^^^feffi^^H). Their exterior is hard, 
while the interior is soft and sweet. Owing to their durability I wish to 
present them to you as a gift.' " 3 It is worthy of note, that, while the 
walnut is said in this text to hail from the Western K'iah, the term 
hu t'ao (not kHan Vao) is employed; so that we may infer that the intro- 
duction of the fruit from the Hu preceded in time the introduction 
from the K'iah. It is manifest also that in this narrative the walnut 
appears as a novelty. 

The Tibetan name of the walnut in general corresponds to a type 
tar-ka, as pronounced in Central Tibetan, written star-ka, star-ga, 
and dar-sga} The last-named spelling is given in the Polyglot Dic- 
tionary of K'ien-luh, 5 also in Jaschke's Tibetan Dictionary. The element 
ka or ga is not the well-known suffix used in connection with nouns, 6 
but is an independent base with the meaning " walnut," as evidenced 
by Kanaurl ka ("walnut"). 7 The various modes of writing lead to a 
restitution *tar, dar, d'ar (with aspirate sonant). This word is found 
also in an Iranian dialect of the Pamir: in Waxi the walnut is called 

1 He died in a.d. 328. His biography is in the Tsin Su, Ch. 100, p. 9. See also 
L. Wieger, Textes historiques, p. 1086. 

2 Literally, "flying stalk of grain." Bretschneider and Stuart do not mention 
this plant. Dr. T. Tanaka, assistant in the Bureau of Plant Industry, Department 
of Agriculture, Washington, tells me that fei-San is a synonyme of the fingered citrus 
(/m Sou kan / fj$j ^ tfl*, Citrus chirocarpus). He found this statement in the Honzo 
kdmoku keimo (Ch. 26, p. 18, ed. 1847) by Ono Ranzan, who on his part quotes the 
T'un ya $§ #ft by Fan I-ci. 

3 The Tai p'in yu Ian reads % VX §? $C JrJL ^ M The Tan Sun pai k'un 
leu Vie and the T'u S'u tsi ten, however, have ^fimi^lif^JcJiJL^Jli "their 
substance resembles the ancient sages, and I wish to present them," — apparently a 
corruption of the text. 

4 W. W. Rockhill (Diary of a Journey through Mongolia and Tibet, p. 340) 
gives taga as pronunciation in eastern Tibet. J. D. Hooker (Himalayan Journals, 
p. 237) offers taga-Hn (Sin, "tree") as Bhutia name. 

6 Ch. 28, p. 55. 

6 Schiefner, Melanges asiatiques, Vol. I, pp. 380-382. 

7 Given both by T. R. Joshi (Grammar and Dictionary of the Kanawari Lan- 
guage, p. 80) and T. G. Bailey (Kanauri-English Vocabulary, Journal Royal As. 
Soc, 191 1, p. 332). Bailey adds to the word also the botanical term Juglans regia. 
The same author, further, gives a word ge as meaning "kernel of walnut; edible part 
of Pinus gerardiana" ; while Joshi (p. 67) explains the same word as the "wild 
chestnut." Thus it seems that ge, ka, originally referred to an indigenous wild-grow- 
ing fruit, and subsequently was transferred to the cultivated walnut. 

The Walnut 261 

tar. 1 This apparently is a loan-word received from the Tibetan, for in 
Sariqoli and other Pamir dialects we find the Iranian word ghdz? 
Tarka is a genuine Tibetan word relating to the indigenous walnut, 
wild and cultivated, of Tibetan regions. In view of this state of affairs, 
it is certainly possible that the Chinese, in the beginning of the fourth 
century or somewhat earlier, received walnuts and their seeds also 
from Tibetan tribes, which resulted in the name KHan t'ao. The 
Lepcha of Sikkim are acquainted with the walnut, for which they have 
an indigenous term, k6l-p6t, and one of their villages is even called 
"Walnut-Tree Foundation" (K61-bah). 3 

G. Watt 4 informs us that the walnut-tree occurs wild and cultivated 
in the temperate Himalaya and Western Tibet, from Kashmir and 
Nubra eastwards. W. Roxburgh 5 says about Juglans regia, "A native 
of the mountainous countries immediately to the north and north-east 
of Hindustan, on the plains of Bengal it grows pretty well, but is not 
fruitful there." Another species of the same genus, J. plerococca Roxb., 
is indigenous in the vast forests which cover the hills to the north and 
east of the province of Silhet, the bark being employed for tanning, while 
/. regia is enlisted among the oil-yielding products. 6 J. D. Hooker 7 
is authority for the information that the walnut occurs wild in Sikkim, 
and is cultivated in Bhutan, where also Captain Turner 8 found it 
growing in abundance. Kirkpatrick 9 met it in Nepal. In Burma it 
grows in the Ava Hills. In the Shan states east of Ava grows another 
species of Juglans, with smaller, almost globose, quite smooth nuts, 
but nothing is known about the tree itself. 10 

The Tibetans certainly cultivate the walnut and appreciate it 

1 R. B. Shaw, On the Ghalchah Languages {Journal As. Soc. Bengal, 1876, 
p. 267), writes the word tor. A. Hujler (The Languages Spoken in the Western 
Pamir, p. 36, Copenhagen, 1912) writes tar, explaining the letter a as a "dark deep a, 
as in the French pas." 

2 W. Tomaschek (Pamirdialekte, p. 790) has expressed the opinion that WaxJ 
tor, as he writes, is hardly related to Tibetan star-ga; this is not correct. 

8 G. Mainwaring, Dictionary of the Lepcha Language, p. 30. 

4 Dictionary of the Economic Products of India, Vol. IV, p. 550. 

8 Flora Indica, p. 670. 

6 N. G. Mukerji, Handbook of Indian Agriculture, p. 233. 

7 Himalayan Journals, p. 235; also Risley, Gazetteer of Sikkim, p. 92 (compare 
Darwin, Variation of Animals and Plants under Domestication, Vol. I, p. 445). 

8 Account of an Embassy to the Court of the Teshoo Lama, p. 273. Also Eden 
and Pemberton (Political Missions to Bootan, p. 198, Calcutta, 1895) mention 
the walnut in Bhutan. 

9 Account of Nepaul, p. 81. 

10 S. Kurz, Forest Flora of British Burma, Vol. II, p. 490 (Calcutta, 1877). 

262 Sino-Iranica 

much. The tree is found everywhere in eastern Tibet where horti- 
culture is possible, and among the Tibetan tribes settled on the soil 
of Se-c'wan Province. W. W. Rockhill 1 even mentions that in the 
Ba-t'an region barley and walnuts are used in lieu of subsidiary coinage. 
Lieut.-Col. Waddell 1 makes two references to cultivated walnut-trees 
in Central Tibet. The Chinese authors mention "Tibetan walnuts" 
as products of the Lhasa district. 3 

While the Can-K'ien tradition is devoid of historical value, and 
must be discarded as an historical fact, yet it is interesting from a 
psychological point of view; for it shows at least that, at the time when 
this fiction sprang into existence, the Chinese were under the impression 
that the walnut was not an indigenous tree, but imported from abroad. 
An autochthonous plant could not have been made the object of such a 
legend. A direct reference to the introduction of the cultivated walnut 
with an exact date is not extant in Chinese records, but the fact of such 
an introduction cannot reasonably be called into doubt. It is supported 
not only by the terms hu Vao and k'ian Vao ("peach of the Hu," "peach 
of the K'ian"), but also by the circumstantial evidence that in times 
of antiquity, and even under the Han, no mention is made of the 
walnut. True it is, it is mentioned in the Kin kwei yao lio of the second 
century; but, as stated, this may bean interpolation. 4 Of all the data 
relating to this fruit, there is only one that may have a faint chance to 
be referred to the Han period, but even this possibility is very slight. 
In the Si kin tsa ki Hf M $H ftl 5 it is said that in the gardens of the 
San-lin Park Jt $fc #B of the Han emperors there were walnuts which 
had come from the Western Regions or Central Asia. The Si kin tsa ki, 
however, is the work of Wu Kun £k %, who lived in the sixth century 
a.d., 8 and cannot be regarded as a pure source for tracing the culture 
of the Han. It is not difficult to see how this tradition arose. When the 
San-lin Park was established, the high dignitaries of the empire were 
called upon to contribute famed fruits and extraordinary trees of distant 
lands. We know that after the conquest of Nan-yiie in in B.C. the 
Emperor Wu ordered southern products, like oranges, areca-nuts, 

1 Diary of a Journey through Mongolia and Tibet, p. 347. 

* Lhasa and its Mysteries, pp. 307, 315. See also N. V. KtfNER, Description of 
Tibet (in Russian), Vol. I, pt. 2, p. 137. 

* Rockhill, Journal Royal As. Soc, 1891, p. 273. 

4 Above, p. 205. Can Ki says or is made to say, "Walnuts must not be eaten in 
large quantity, for they rouse mucus and cause man to drink" (Ch. c, p. 27). 

6 Ch. 1, p. 6 (ed. of Han Wei ts'un lu). 

8 Wylie, Notes on Chinese Literature, p. 189; and Chavannes, Voung Pot, 
1906, p. 102. 

The Walnut 263 

lun nan, It-Si, etc., to be brought to the capital C'an-nan, and to be 
planted in the Fu-li Palace £fc %b »lf , founded in commemoration of the 
conquest of Nan-yue, whereupon many gardeners lost their lives when 
the crops of the li-U proved a failure. 1 Several of his palaces were named 
for the fruits cultivated around them: thus there were a Grape-Palace 
and a Pear-Palace. Hence the thought that in this exposition of foreign 
fruits the walnut should not be wanting, easily impressed itself on the 
mind of a subsequent writer. Wu Kun may also have had knowledge 
of the Can-K'ien tradition of the Po wu U, and thus believed himself 
consistent in ascribing walnuts to the Han palaces. Despite his ana- 
chronism, it is interesting to note Wu Kun's opinion that the walnut 
came from Central Asia or Turkistan. 

It is not probable that the walnut was generally known in China 
earlier than the fourth century a.d., under the Eastern Tsin M U 
dynasty (2 65-41 o). 2 In the Tsin kun ko min If *$£ 18 ^, a description 
of the palaces of the Tsin emperors, written during that dynasty, 3 it is 
stated that there were eighty-four walnut-trees in the Hwa-lin Park 

1 The palace Fu-li was named for the li-li -j$ $£ (see Sanfu hwan t'u H $§ ^C 
gS . Ch. 3, p. 9 b, ed. of Han Wei ts'un lu). 

1 Bretschneider (Bot. Sin., pt. I, p. 39) asserts that Juglans regia figures 
among the plants mentioned passingly in the Nan fan ts'ao mu Swan by Ki Han 
|f &, a minister of state under the Emperor Hui j§ of the Tsin dynasty 
(a.d. 290-306) . He does not give any particulars. There are only two allusions to the 
walnut, that I am able to trace in this work: in the description of the coco-nut, 
the taste of this fruit is likened to that of the walnut; and the flavor of the "stone 
chestnut" {li-li ^EJ ££, Aleurites triloba) is compared with that of the same fruit. 
We know at present that the book in question contains interpolations of later date 
(see L. Aurousseau, Bull, de I'Ecole franqaise, Vol. XIV, 1914, p. 10); but to these 
the incidental mention of the walnut does not necessarily belong, as Ki Han lived 
under the Tsin. It is likewise of interest that the walnut is not dealt with as a special 
item in the Ts'i min yao lu, a work on husbandry and economic botany, written by 
Kia Se-niu |a£ JB> §.&> of the Hou Wei dynasty (a.d. 386-534) ; see the enumeration 
of plants described in this book in Bretschneider {op. cit., p. 78). In this case, the 
omission does not mean that the tree was unknown to the author, but it means only 
that it had then not attained any large economic importance. It had reached the 
palace-gardens, but not the people. In fact, Kia Se-niu, at least in one passage 
(Ch. 10, p. 48 b, ed. 1896), incidentally mentions the walnut in a quotation from the 
Kiao lou ki ^ >}\\ |2 by Liu Hin-k'i £lj /ft J0, where it is said, "The white yuan 
tree £j K«l [evidently = |$|<] is ten feet high, its fruits being sweeter and finer 
than walnuts $J $|j." As the Kiao lou ki is a work relating to the products of 
Annam, it is curious, of course, that it should allude to the cultivated walnut, which 
is almost absent in southern China and Annam; thus it is possible that this clause 
may be an interpolation, but possibly it is not. The fact that the same work like- 
wise contains the tradition connecting the walnut with Can K'ien has been pointed 
out above. The tree pai yuan is mentioned again in the Pen ts'ao kan mu li i (Ch. 8, 
p. 23), where elaborate rules for the medicinal employment of the fruit are given. 

• Bretschneider, Bot. Sin., pt. 1, p. 202, No. 945. 

264 Sino-Iranica 

3$i # 13. * Another allusion to the walnut relative to the period Hien-ho 
(a.d. 326-335) has been noted above (p. 259). There is, further, a refer- 
ence to the fruit in the history of Su ID , when, after the death of Li Hiun 
3p M in a.d. 334, Han Pao & J& from Fu-fuh #v $& in Sen-si 
was appointed Grand Tutor (t'ai fu zk $§•) of his son Li K'i ^ $J, and 
asked the latter to grant him seeds for the planting of walnut-trees, 
which, on account of his advanced age, he was anxious to have in his 
garden. 2 

During the third or fourth century, the Chinese knew also that 
walnuts grew in the Hellenistic Orient. "In Ta Ts'in there are jujubes, 
jasmine, and walnuts," it is stated in the Wu H wai kwo li ^ Bvf ^ 
@8 jS ("Memoirs of Foreign Countries at the time of the Wu"). 3 

The Kwan li )0( jS by Kwo Yi-kuh IP H # 4 contains the following 
account: "The walnuts of C'en-ts'ah W. M 5 have a thin shell and a 
large kernel; those of Yin-p'ih ^ Z P 6 are large, but their shells are brittle, 
and, when quickly pinched, will break." 7 

Coming to the T'ang period, we encounter a description of the 
walnut in the Yu yah tsa tsu @ $1 H 21, written about a.d. 860, 8 from 
which the fact may be gleaned that the fruit was then much cultivated 

1 T'ai p'in yti Ian, I.e. 

2 This story is contained in the Kwan wu hin ki ]§( 3t ff IE (according to 
Bretschneider, a work of the Sung literature). As the text is embodied in the 
T'ai p'in yti Ian, it must have been extant prior to a.d. 983, the date of Li Fan's 

3 Presumably identical with the Wu Si wai kwo Swan noted by Pelliot (Bull, de 
I'Ecole jrancaise, Vol. IV, p. 270) as containing information secured by the mission 
of K'an T'ai in the first part of the third century a.d. Cf. also Journal asiatique, 
191 8, II, p. 24. The Min H ascribes walnuts to Ormuz (Bretschneider, Notices 
of the Mediaeval Geography, p. 294). 

4 This work is anterior to the year a.d. 527, as it is cited in the Swi kin lu of 
Li Tao-yuan, who died in that year. Kwo Yi-kuh is supposed to have lived under 
the Tsin (a.d. 265-419). Cf. Pelliot, Bull, de I'Ecole jrancaise, Vol. IV, p. 412. 

6 Now the district of Pao-ki in the prefecture of Fuh-sian, Sen-si Province. 

6 At the time of the Han period, Yin-p'ih was the name for the present prefec- 
ture of Lun-han ft ^ in the province of Se-2'wan. There was also a locality of the 
same name in the prefecture of Kiai in the province of Kan-su, inhabited by the Ti, 
a Tibetan tribe (Chavannes, T'oung Pao, 1905, p. 525). 

7 T'ai p'in yti Ian, I. c; Ko li kin yuan, Ch. 76, p. 5; Ci wu min H t'u k'ao, I. c. 
This text is cited also by Su Sun in his T'u kin pen ts'ao. The earliest quotation 
that I can trace of it occurs in the Pei hu lu, written by Twan Kuh-lu about a.d. 
875 (Ch. 3, p. 4 b, ed. of Lu Sin-yuan), where, however, only the last clause in regard 
to the walnuts of Yin-p'in is given (see below, p. 268). 

8 Pelliot, T'oung Pao, 1912, p. 375. The text is in the T'u Su tsi I'en and 
Ci wu min si t'u k'ao (I. c). I cannot trace it in the edition of the Yu yan tsa tsu in 
the Tsin tai pi su or Pai hai. 

The Walnut 265 

in the northern part of China (^b ~JS %> M £,), — a statement repeated 
in the K'ai-pao pen ts'ao. The Yu yah tsa tsu, which is well informed 
on the cultivated plants of Western and Central Asia, does not contain 
the tradition relating to Can K'ien, but, on the other hand, does not 
speak of the tree as a novel introduction, nor does it explain its name. 
It begins by saying that "the kernel of the walnut is styled 'toad' 
ha-mo i|t ^. " l 

Mon Sen j£ I5fe, who in the second half of the seventh century wrote 
the $i liao pen ts'ao, 2 warns people from excessive indulgence in walnuts 
as being injurious to health. 3 The T'ai p'ih hwan yu ki ;fc *r &t -f' Hti, 
by Yo Si f^ it. (published during the period T'ai-p'ih, a.d. 976-981), 
mentions the walnut as being cultivated in the prefecture of Fun-sian 
B* ffli in Sen-si Province, and in Kiah cbu 1& 'M in San-si Province. 4 

According to the Pen ts'ao kah mu, the term hu t'ao first appears in 
the Pen ts'ao of the K'ai-pao period (968-976) of the Sung dynasty, 
written by Ma Ci E$ iS; that is to say, the plant or its fruit was then 
officially sanctioned and received into the pharmacopoeia for the first 
time. We have seen that it was certainly known prior to that date. 
K'ou Tsuh-si 7cl ^ Ifil, in his Pen ts'ao yen i ^ ^ $T i§ of 1116, 5 has a 
notice on the medicinal application of the fruit. 

It is possible also to trace in general the route which the walnut has 
taken in its migration into China. It entered from Turkistan into 
Kan-su Province, as stated by Su Sun (see above, p. 258), and gradually 
spread first into Sen-si, and thence into the eastern provinces, but always 
remained restricted to the northern part of the country. Su Sun ex- 
pressly says that walnuts do not occur in the south, but only in the 
north, being plentiful in Sen-si and Lo-yan (Ho-nan Province), while 
those grown in K'ai-fun (Pien cou^ 'M) were not of good quality. In the 
south only a wild-growing variety was known, which is discussed 
below. Wan Si-mou 3: 1fr ^, a native of Kian-su, who died in 1591, 
states in his Kwo su ^ Eft, a treatise on garden-fruits, that "the walnut 
is a northern fruit (pei kwo ^b ^), and thrives in mountains; that it 
is but rarely planted in the south, yet can be cultivated there." Almost 

1 This definition is ascribed to the Ts'ao mu tse ^L ;fv ~F" in the Ko li kin yuan 
(Ch. 76, p. 5); that work was written by Ye Tse-k'i ^ J" -frf in 1378 (Wylie, 
Notes on Chinese Literature, p. 168). 

2 Bretschneider, Bot. Sin., pt. 1, p. 45. 

3 T'an Sun pai k'un leu Vie, Ch. 99, p. 12. 

4 T'ai p'in hwan yu ki, Ch. 30, p. 4; Ch. 47, p. 4 (ed. of Kin-lin Su ku, 1882). 

6 Ch. 18, p. 6 b (ed. of Lu Sin-yuan). 

6 Also J. de Loureiro (Flora cochinchinensis, p. 702) states that the habitat of 
Juglans regia is only in the northern provinces of China. 

*66 Sino-Iranica 

all the district and prefectural gazetteers of Sen-si Province enumerate 
the walnut in the lists of products. The "Gazetteer of San-tun" 1 
mentions walnuts for the prefectures of Ts'i-nan, Yen-cou, and Ts'in- 
cou, the last-named being the best. The Gazetteer of the District of 
Tun-no 3lC ffl 2 in the prefecture of Tai-nan in San-tun reports an 
abundance of walnuts in the river-valleys. An allusion to oil-production 
from walnuts is found in the "Gazetteer of Lu-nan," where it is said, 
"Of all the fruits growing in abundance, there is none comparable to 
the walnut. What is left on the markets is sufficient to supply the needs 
for lamp-oil." 8 Also under the heading "oil," walnut-oil is mentioned 
as a product of this district. 4 

Juglans regia, in its cultivated state, has been traced by our botanists 
in San-tun, Kian-su, Hu-pei, Yiin-nan, and Se-6'wan. 6 Wilson nowhere 
saw trees that could be declared spontaneous, and considers it highly 
improbable that Juglans regia is indigenous to China. His opinion is 
certainly upheld by the results of historical research. 

A wild species (Juglans mandshurica or caihayensis Dode) occurs 
in Manchuria and the Amur region, Ci-li, Hu-pei, Se-5'wan, and Yun- 
nan.' This species is a characteristic tree of the Amur and Usuri val- 
leys. 7 It is known to the Golde under the name kofoa or koloa, to the 
Managir as korfo, to the Gilyak as tiv-alys. The Golde word is of 
ancient date, for we meet it in the ancient language of the Jurci, JuSen, 
or Niuci in the form xu!>u s and in Manchu as xosixa. The great antiquity 
of this word is pointed out by the allied Mongol word xusiga. The 
whole series originally applies to the wild and indigenous species, 

1 San tun t'un H, Ch. 9, p. 15. 

• Ch. 2, p. 32 (1829). 

8 Quotation from Lu-nan U %% f^J j^, in the San cou tsun U $fj j{\ |§ jg 
(General Gazetteer of San-cou), 1744, Ch. 8, p. 3. 

4 Ibid., Ch. 8, p. 9. Oil was formerly obtained from walnuts in France both 
for use at table and for varnishing and burning in lamps, also as a medicine sup- 
posed to possess vermifuge properties (Ainslie, Materia Indica, Vol. I, p. 464). 

6 See particularly C. S. Sargent, Plantae Wilsonianae, Vol. Ill, pp. 184-185 
(1916). J. Anderson (Report on the Expedition to Western Yunan, p. 93, Calcutta, 
1 871) mentions walnuts as product of Yiin-nan. According to the Tien hai yu hen 
li (Ch. 10, p. 1 b; above, p. 228), the best walnuts with thin shells grow on the Yan-pi 
or Yah-p'ei River tH ifo Ql of Yiin-nan. 

• Forbes and Hemsley, Journal of the Linnean Society, Botany, Vol. XXVI, 
p. 493; Sargent, op. cit., pp. 185 et seq. J. de Loureiro (Flora cochinchinensis, 
p. 702), writing in 1788, has a species Juglans camirium (Annamese dedu lai) "habitat 
agrestis cultaque in Cochinchina;" and a Juglans catappa (Annamese cdy mo cua) 
''habitat in sylvis Cochinchinae montanis." 

T Grum-Grzimailo, Description of the Amur Province (in Russian), p. 313. 

• W. Grube, Schrift und Sprache der JuCen, p. 93. 

The Walnut 267 

Juglans mandshurica. Manchu xdsixa designates the tree, while its 
fruit is called xdwalama or xdwalame usixa {-ixa being a frequent ter- 
mination in the names of plants and fruits). The cultivated walnut is 
styled mase. 1 One of the earliest explorers of the Amur territory, the 
Cossack chieftain Poyarkov, who reached the Amur in 1644, reported 
that walnuts and hazel-nuts were cultivated by the Daur or Dahur on 
the Dseya and Amur. 1 

The same species is known to the aboriginal tribes of Yun-nan. 
The Pa-yi and San style its fruit twai; 3 the Nyi Lo-lo, se-mi-ma; the Ahi 
Lo-lo, sa-mi. The Cun-kia of Kwei-cou call it dsao; the Ya-c'io Miao, 
U or H] the Hwa Miao, klaeo; while other Miao tribes have the Chinese 
loan-word he-dao.* 

The wild walnut has not remained unknown to the Chinese, and it 
is curious that it is designated San hu t'ao tfj #§ $iS, the term San ("moun- 
tain") referring to wild-growing plants. The "wild Iranian peach" 
is a sort of linguistic anomaly. It is demonstrated by this term that 
the wild indigenous species was discovered and named by the Chinese 
only in times posterior to the introduction of the cultivated variety; and 
that the latter, being introduced from abroad, was not derived from the 
wild-growing species. The case is identical with that of the wild alfalfas 
and vines. C'en Hao-tse, who wrote a treatise on flowers in 1688, 8 
determines the difference between the cultivated and wild varieties 
thus: the former has a thin shell, abundant meat, and is easy to break; 6 
the latter has a thick and hard shell, which must be cracked with a 
hammer, and occurs in Yen and Ts'i (Ci-li and San-tun). This observa- 

1 K'ien-luh's Polyglot Dictionary, Ch. 28, p. 55. 

* L. v. Schrenck, Reisen und Forschungen im Amur-Lande, Vol. Ill, p. 160. 
3 F. W. K. MttLLER, Toung Pao, Vol. Ill, 1892, p. 26. 

* S. R. Clarke, Tribes in South- West China, p. 312. 
6 Hwa kin, Ch. 3, p. 49 b. 

* According to the Ci wu min Si t'u k'ao (Ch. 31, p. 3 b), the walnuts with thin 
shells grow only in the prefecture of Yun-p'in fo ^P in Ci-li, being styled lu tan 
ho t'ao J& U $% $j} In C'an-li, which belongs to this prefecture, these nuts have 
been observed by F. N. Meyer (Agricultural Explorations in the Orchards of China, 
p. 51), who states, "Some trees produce small hard-shelled nuts of poor flavor, while 
others bear fine large nuts, with a really fine flavor, and having shells so thin that 
they can be cracked with the fingers like peanuts. Between these extremes one finds 
many gradations in hardness of shell, size, and flavor." "In England the walnut 
presents considerable differences, in the shape of the fruit, in the thickness of the 
husk, and in the thinness of the shell; this latter quality has given rise to a variety 
called the thin-shelled, which is valuable, but suffers from the attacks of titmice" 
(Darwin, Variation of Animals and Plants under Domestication, Vol. I, p. 445). 
A variety of walnut with thin shells grows on the Greek Island Paros (T. v. Held- 
eeich, Nutzpflanzen Griechenlands, p. 59). 

268 Sino-Iranica 

tion is quite to the point; the shell of the walnut gradually became more 
refined under the influence of cultivation. 

The earliest texts alluding to the wild walnut are not older than 
the T'ang period. The Pet hu lu At P £^, written by Twan Kun-lu 
H & $& about a.d. 875, 1 contains the following text concerning a wild 
walnut growing in the mountains of southern China: — 

"The wild walnut has a thick shell and a flat bottom i£ Z P. In 
appearance it resembles the areca-nut. As to size, it is as large as a 
bundle of betel-leaves. 2 As to taste, it comes near the walnuts of 
Yin-p'ih 3 and Lo-yu, but is different from these, inasmuch as it has a 
fragrance like apricot extract. This fragrance, however, does not last 
long, but will soon vanish. The Kwan ci says that the walnuts of Yin- 
p'in have brittle shells, and that, when quickly pinched, the back of 
the kernel will break. Liu Si-luh W i& P^, in his Sie lo yu yuan M ^ 
M. ^E, remarks, with reference to the term hu t'ao, that the Hu take to 
flight like rams, 4 and that walnuts therefore are prophets of auspicious 
omens. Ceh K'ien lift i^ 5 says that the wild walnut has no glumelle; 
it can be made into a seal by grinding off the nut for this purpose. 
Judging from these data, it may be stated that this is not the walnut 
occurring in the mountains of the south." 6 

The Lin piao lui^k &$&=&, by Liu Sun M 'Nd of the T'ang period, 7 
who lived under the reign of the Emperor Cao Tsuh (a.d. 889-904), 
contains the following information on a wild walnut : — 

"The slanting or glandular walnut {pHen ho Vao M W. $li) is pro- 
duced in the country Can-pi t^ =P. 8 Its kernel cannot be eaten. The 

1 Cf. Pelliot, Bull, de VEcolefrancaise, Vol. IX, p. 223. 

2 Fu-liu, usually written ^ •©, is first mentioned in the Wu lu ti li ci ^ ^ j$J 
?S iS by Can Pu 3H ^] of the third or beginning of the fourth century (see Ts'i 
min yao Su, Ch. 10, p. 32). It refers to Piper betle (Bretschneider, Chinese Recorder, 
Vol. Ill, 1 87 1, p. 264; C. Imbault-Huart, Le b£tel, T'oung Pao, Vol. V, 1894, 
p. 313). The Chinese name is a transcription corresponding to Old Annamese 
bldu; M£sdn, Uy-16, and Hung plu; Khmer tn-luiv, Stien m-lu, Bahnar bo-lou, Kha 
b-lu ("betel"). 

3 See above, p. 264. 

4 A jocular interpretation by punning Vao $|j upon t'ao ^fe (both in the same 

5 Author of the lost Hu pen ts'ao $J ^ ^ (Bretschneider, Bot. Sin., pt. 1, 
p. 45). He appears to have been the first who drew attention to the wild walnut. 
His work is repeatedly quoted in the Pet hu lu. 

6 Pel hu lu, Ch. 3, p. 4 b (ed. of Lu Sin-yuan). 

7 Ch. B, p. 5 (ed. of Wu yin tien). 

8 The two characters are wrongly inverted in the text of the work. In the text 
of the Pei hu lu that follows, the name of this country is given in the form Can-pei 
t^ ^.. From the mention of the Malayan Po-se in the same text, it follows that 

The Walnut 269 

Hu $§ people gather these nuts in abundance, and send them to the 
Chinese officials, designating them as curiosities ^ H. As to their 
shape, they are thin and pointed; the head is slanting like a sparrow's 
beak. If broken and eaten, the kernel has a bitter taste resembling that 
of the pine-seeds of Sin-ra $f M %k -J 1 . 1 Being hot by nature, they are 
employed as medicine, and do not differ from the kernels of northern 

The Pet hu lu 2 likewise mentions the same variety of glandular wal- 
nut (pHen ho-Vao) as growing in the country Can-pei £ ^, shaped 
like the crescent of the moon, gathered and eaten by the Po-se, 3 having 
a very fine fragrance, stronger than the peach-kernels of China, but of 
the same effect in the healing of disease. 

The species here described may be identical with Juglans catha- 
yensis, called the Chinese butternut, usually a bush, but in moist 
woods forming a tree from twelve to fifteen metres tall; but I do not 
know that this plant occurs in any Malayan region. With reference to 
Can-pi, however, it may be identical with the fruit of Canarium com- 
mune (family Burseraceae), called in Malayan kanari, in Javanese kenari. 
J. Crawfurd, 4 who was not yet able to identify this tree, offers the 
following remarks: "Of all the productions of the Archipelago the one 
which yields the finest edible oil is the kanari. This is a large handsome 
tree, which yields a nut of an oblong shape nearly of the size of a walnut. 
The kernel is as delicate as that of a filbert, and abounds in oil. This 

Can-pi is a Malayan territory probably to be located on Sumatra. For this reason 
I am inclined to think that Can-pi r!j ^ is identical with Can-pei 41 W^ '< that is. 
Jambi, the capital of eastern Sumatra (Hirth and Rockhill, Chau Ju-kua, pp. 65, 
66; see further Groeneveldt, Notes on the Malay Archipelago, pp. 188, 196; and 
Gerini, Researches on Ptolemy's Geography, p. 565; Lin wai tai la, Ch. 2, p. 12). 
From a phonetic point of view, however, the transcription r^j ^, made in the 
T'ang period, represents the ancient sounds *£an-pit, and would presuppose an 
original of the form *cambit, 2ambir, or jambir, whereas ^L is without a final con- 
sonant. The country Can-pei is first mentioned under the year a.d. 852 (^ 4* sixth 
year), when Wu-sie-ho ^J ffi JH& and six men from there came to the Chinese Court 
with a tribute of local products (T'ai p'in hwan yu ki, Ch. 177, p. 15 b). A second 
embassy is on record in 871 (Pelliot, Bull, de I'Ecole frangaise, Vol. IV, p. 347). 

1 Pinus koraiensis Sieb. et Zucc. (J. Matsumura, Shokubutsu mei-i, pp. 266-267, 
ed. 1915), in Japanese losen-matsu ("Korean pine"); see also Stuart, Chinese 
Materia Medica, p. 333. Sin-ra (Japanese Sin-ra, Siraki) is the name of the ancient 
kingdom of Silla, in the northern part of Korea. 

2 Ch. 3, p. 5 (ed. of Lu Sin-yuan). 

3 Wi $ft certainly is here not Persia, for the Pet hu lu deals with the products 
of Kwari-tuh, Annam, and the countries south of China (Pelliot, Bull, de I'Ecole 
francaise, Vol. IX, p. 223). See below, p. 468. The Pel hu lu has presumably served 
as the source for the text of the Lin piao lu i, quoted above. 

* History of the Indian Archipelago, Vol. I, p. 383. 

tyo Sino-Iranica 

is one of the most useful trees of the countries where it grows. The 
nuts are either smoked and dried for use, or the oil is expressed from 
them in their recent state. The oil is used for all culinary purposes, 
and is more palatable and finer than that of the coconut. The kernels, 
mixed up with a little sago meal, are made into cakes and eaten as 
bread. The kanari is a native of the same country with the sago tree, 
and is not found to the westward. In Celebes and Java it has been 
introduced in modern times through the medium of traffic." 

The Yu yah tsa tsu 1 speaks of a man hu t'ao M fi§ $& as "growing 
in the kingdom of Nan-Cao ^j la in Yun-nan; it is as large as a flat 
conch, and has two shells of equal size; its taste is like that of the 
cultivated walnut. It is styled also 'creeper in the land of the Man' 
(Man luh t'eh-tse S$ 't'llfF')-" It will be remembered that Twan 
C'eh-§i, the author of this work, describes also the cultivated walnut 
(p. 264). 

The T'ai p'in yii Ian contains another text attributed to the Lin 
piao lu i relating to a wild walnut, which, however, is not extant in the 
edition of this work published in the collection Wu yih tien in 1775. 
This text is as follows: "The large walnut has a thick and firm shell. 
It is larger than that of the areca-nut. 2 It has much meat, but little 
glumelle. It does not resemble the nuts found in northern China. It 
must be broken with an axe or hammer. The shell, when evenly 
smoothed over the bottom, is occasionally made into a seal, for the 
crooked structure of the shell {ko Rtf) resembles the seal characters." 8 

In the Lin wai tai ta ^ ft ft ^, 4 written by Cou K'u-fei ffl £ $r 
in 1 1 78, mention is made, among the plants of southern China and 
Tonking, of a "stone walnut (U hu Vao ^ $] $&), which is like stone, 
has hardly any meat, and tastes like the walnut of the north." Again, 
a wild species is involved here. I have not found the term H hu Vao in 
any other author. 

The various names employed by the T'ang writers for the wild 

1 Ch. 19, p. 9 b (ed. of Tsin tai pi Su); or Ch. 19, p. 9 a (ed. of Pai hai). 

1 This sentence, as well as the first, agrees with the definition given by the Pei 
hu lu with reference to a wild walnut (above, p. 268). 

1 T'ai p'in yii Ian, Ch. 971, p. 8 b. The same text is cited by the Pen ts'ao kan 
mu and the Ko li kin yuan (Ch. 76, p. 5 b), which offer the reading San hu t'ao [I] 
§J ^ ("wild walnut") instead of "large walnut." The Kwan k'iinfan p'u (Ch. 58, 
p. 26) also has arranged this text under the general heading "wild walnut." The 
Pen ts'ao kan mu opens it with the sentence, "In the southern regions there is a wild 
walnut." The restriction to South China follows also from the text as given in the 
T'ai p'in yii Ian. 

* Ch. 8, p. 10 b (ed. of Ci pu tsu lai ts'un Su). 

The Walnut 971 

varieties {p'ien hu t'ao, San hu t'ao, man hu t'ao, ta hu t'ao), combined 
with the fact that two authors describe both the varieties p'ien and 
San, raise the question whether this nomenclature does not refer to 
different plants, and whether, aside from the wild walnut, other nuts 
may not also be included in this group. In this respect it is of interest 
to note that the hickory, recently discovered in Ce-kian by F. N. 
Meyer, and determined by Sargent 1 under the name Carya cathayensis, 
is said by Meyer to be called shan-gho-to in the colloquial language; 
and this evidently is identical with our San hu t'ao. This certainly does 
not mean that this term refers exclusively to the hickory, but only 
that locally the hickory falls also within the category of San hu t'ao. 
The distribution of the hickory over China is not yet known, and the 
descriptions we have of San hu t'ao do not refer to Ce-kiah. 

In the P'an San U M ill A>, a description of the P'an mountains, 8 
the term San ho t'ao is given as a synonyme for the bark of Catalpa 
bungei (ts'iu p'i tfk $t), which is gathered on this mountain for 
medicinal purposes, — presumably because the structure of this bark 
bears some superficial resemblance to that of a walnut. Wild walnuts, 
further, are mentioned as growing on Mount Si fu 2un j§ ^ H \U , 
forming part of the Ma-ku Mountains M. ft£ \U situated in Fu-cou 
flU 'H\ in the prefecture of Kien-C'an M. H /fr, Kiah-si Province. 8 

While the cultivated walnut was known in China during the fourth 
century under the Tsin dynasty, the wild species indigenous to south- 
ern China was brought to the attention of scholars only several cen- 
turies later, toward the close of the T'ang period. This case furnishes 
an excellent object-lesson, in that it reveals the fallacies to which 
botanists and others are only too frequently subject in drawing con- 
clusions from mere botanical evidence as to cultivated plants. The 
favorite argumentation is, that if, in a certain region, a wild and a 
corresponding cultivated species co-exist, the cultivated species is simply 
supposed to have been derived from the wild congener. This is a de- 
ceptive conclusion. The walnut (as well as the vine) of China offers a 

1 Plantae Wilsonianae, Vol. Ill, p. 187. 

* Ch. 15, p. 2 b, of the edition published in 1755 by order of K'ien-luh. The 
P'an San is situated three or four days' journey east of Peking, in the province of 
Ci-li, the summit being crowned by an interesting Buddhist temple, and there being 
an imperial travelling-station at its foot. It was visited by me in September, 1901. 
P. N. Meyer (Agricultural Explorations in the Orchards of China, p. 52) says that 
in the Pangshan district east of Peking one may still find a few specimens of the real 
wild walnut growing in ravines among large bowlders in the mountains. 

' Ma-ku San U (Ch. 3, p. 6 b), written by members of the family Hwah ^f, and 
published in 1866 by the Tun t'ien su wu t^I ^ 1§f M.- These mountains contain 
thirty-six caves dedicated to the Taoist goddess Ma-ku. 

272 Sino-Iranica 

specific case apt to teach just the opposite: a wild walnut (probably in 
several species) is indigenous to China, nevertheless the species culti- 
vated in this area did not spring from domestic material, but from 
seeds imported from Iranian and Tibetan regions of Central Asia. 
The botanical dogma has been hurled against many deductions of 
Hehn: botanists proclaimed that vine, fig, laurel, and myrtle have been 
indigenous to Greece and Italy in a wild state since time immemorial; 
likewise pomegranate, cypress, and plantain on the Aegean Islands 
and in Greece; hence it was inferred that also the cultivations of these 
plants must have been indigenous, and could not have been introduced 
from the Orient, as insisted on by Hehn. This is nothing but a sophism: 
the botanists still owe us the proof that the cultivated species were 
really derived from indigenous stock. A species may indeed be indige- 
nous to a certain locality; and yet, as brought about by historical 
inter-relations of the peoples, the same or a similar species in the 
cultivated state may have been introduced from an outside quarter. 
It is only by painstaking historical research that the history of culti- 
vated plants can be exactly determined. Engler (above, p. 258) doubts 
the occurrence of the wild walnut in China, because a cultivated species 
was introduced there from Tibet ! It is plain now where such logic will 
lead us. Wilson deserves a place of honor among botanists, for, after 
close study of the subject in China, he recognized that "it is highly 
improbable that Juglans regia is indigenous to China." 

With reference to the walnut, conditions are the same in China as 
in the Mediterranean region: there also Juglans regia grows spontane- 
ously; still better, cultivated varieties reached the Greeks from Persia; 
the Greeks handed these on to the Romans; the Romans transplanted 
them to Gallia and Germania. Juglans regia occupies an extensive 
natural area throughout the temperate zone, stretching from the 
Mediterranean through Iran and the Himalaya as far as southern China 
and the Chinese maritime provinces. Despite this natural distribution, 
the fact remains that Iran has been the home and the centre of the 
best-cultivated varieties, and has transmitted these to Greece, to India, 
to Central Asia, and to China. 

Dr. T. Tanaka has been good enough to furnish the following infor- 
mation, extracted from Japanese literature, in regard to the walnut. 

"Translation of the notice on ko-td {kurumi), 'walnut,' from a 
Japanese herbal Yamato honzd ^C % P if ^, by Kaibara Ekken j| W* 
^ $f (Ch. 10, p. 23), published in 1709. 

"Kurumi t$ $iS {koto). There are three sorts of walnut. The first 
is called oni-gurumi 3& #] #ii ('devil walnut'). It is round in shape, 

The Walnut 273 

and has a thick, hard skin (shell), difficult to break; it has very little 
meat. In the Honzo (Pen ts'ao, usually referring to the Pen ts'ao kan 
mu) it is called tfj tfi #ii (yama-gurumi, Ian hu t'ao). It is customary 
to open the shell by first baking it a little while in a bed of charcoal, 
and suddenly plunging it in water to cool off; then it is taken out of the 
fire, the shell is struck at the joint so that it is crushed, and the meat can 
be easily removed. The second variety is called hime-gurumi #8! P 
^ * ('demoiselle walnut'), and has a thin shell which is somewhat 
flat in form; it is very easily broken when struck with an iron hammer 
at the joint. It has plenty of meat, is rich in oil, and has a better taste 
than the one mentioned before. The names 'devil' and 'demoiselle' 
are derived from the appearance of the nuts, the one being rough and 
ugly, while the other is beautiful. 

"The third variety, which is believed to have come from Korea, 
has a thin shell, easily cracked, with very little meat, but of the best 
quality. Moh Sen J& Ife (author of the Si liao pen ts'ao Jt ^ ^ 3£, 
second half of the seventh century) says, 'The walnut, when eaten, 
increases the appetite, stimulates the blood-circulation, and makes one 
appear glossy and elegant. It may be considered as a good medicine of 
high merit.' For further details refer to the prescriptions of the Pen 

"Translation of the notice on walnut from the Honzo kdmoku keimo 
(Ch. 25, pp. 26-27) by Ono Ranzan; revised edition by Igu£i Bosi 
of 1847 (first edition 1804). 

11 koto, kurimi (walnut, Juglans regia L., var. sinensis Cas., ex Matsu- 
mura, Shokubutsu Mei-i, ed. 1915, Vol. I, p. 189). 

"Japanese names: to-kurimi ('Chinese walnut'); losen-kurimi 
('Korean walnut'). 

"Chinese synonymes: kaku-kwa (Jibutsu imei) ; linso kyoho (ibid.); 
inpei linhwa (ibid.); kokaku (Jibutsu kon$u); kens' a (ibid.); td$uU 
(Kunmo jikwai). 

"Names for kernels: kama (Roy a taisui-hen). 

"Other names for San hu t'ao: sankakuto (Hokuto-roku); banzai-H 
(Jonan HoU); iii (Kummo jikwai). 

"The real walnut originated in Korea, and is not commonly planted 
in Japan. 

"The leaves are larger than those of onigurumi (giant walnut, 
Juglans sieboldiana Maxim., ex Matsumura, I.e.). The shells are also 
larger, measuring more than 1 sun (1.193 inches) in length, and having 
more striations on the surface. The kernels are also larger, and have 
more folds. 

"The variety commonly planted in our country is onigurumi, the 

a74 Sino-Iranica 

abbreviated name of which is kurumi; local names are ogurumi (Prov- 
ince of Kaga), okkoromi (eastern provinces), and so on. This giant wal- 
nut grows to a large tree. Its leaves are much like those of the lacquer- 
tree (Rhus vernificera DC.) and a little larger; they have finely serrated 
margins. Its new leaves come out in the spring. It flowers in the 

"The flower-clusters resemble chestnut-catkins, but are much 
larger, ranging in length from six to seven sun; they are yellowish white 
and pendulous. A single flower is very small, like that of a chestnut. 
The fruit is peach-shaped and green, but turns black when ripe. The 
shells are very hard and thick, and can be opened by being put on the 
fire for a little while; then insert' a knife in the slit or fissure between the 
shells, which thus break. The kernels are good for human food, and 
are also used for feeding little birds. 

"One species called hime-gurumi ('demoiselle walnut,' Juglans 
cordijormis Maxim., ex Matsumura, I.e.), or me-gurumi ('female wal- 
nut,' from the province of Kaga), has thin shells with fewer furrows, and 
the kernels can easily be taken out. Under the heading lukai (£i-kie, 
explanatory information in the Pen ts'ao), this kind of walnut is de- 
scribed as 'a walnut produced in Cinso (C'en-ts'an, a place in Fuh- 
siah fu, Sen-si, China) with thin shells and many surfaces,' so we call 
it linso-gurumi (Pen-ts'an hu-Vao). 1 This variety is considered the 
best of all yama-gurumi (Ian hu Vao, wild walnuts), because no other 
variety has such saddle-shaped kernels entirely removable from the 

"A species called karasu-gurumi ('crow walnut') is a product of the 
province of E6igo; it has a shell that opens by itself when ripe, and 
looks like a crow's bill when opened, whence it is called 'crow walnut.' 

"Another variety from 0§io-mura village of the Aidzu district is 
called gonroku-gurumi ('Gonroku's walnut'); it has a very small shell 
capable of being used as ojime ('string-fastener of a pouch'). This 
name is taken from the personal name of a man called Anazawa Gon- 
roku, in whose garden this variety originated. It is said that the same 
kind has been found in the province of Kai. 

"A variety found at Nosuo, province of U§Q (Uzen and Ugo), 
is much larger in size, and has thinner shells, easily crushed by hand, 
so that the kernels may be taken out without using any tools. The 
name of this variety is therefore teuU-gurumi ('hand-crushed walnut')." 

The most interesting point in these Japanese notes is presented by 

1 Compare above, p. 264. 

The Walnut 273 

the tradition tracing the cultivated walnut of Japan to Korea. The 
Koreans again have a tradition that walnuts reached them from China 
about fifteen hundred years ago in the days of the Silla Kingdom. 1 
The Korean names for the fruit are derived from the Chinese: ho do 
being the equivalent of hu t'ao, kan do corresponding to k'ian Vao t 
and ha do to ho t'ao. The Geography of the Ming Dynasty states that 
walnuts are a product of Korea. 1 

1 Korea Review, Vol. II, 1902, p. 394. 
* Ta Mi* i t'uA li, Ch. 89 p. 4 b. 


5. A. de Candolle 1 sums up the result of his painstaking investi- 
gation of the diffusion of the pomegranate {Punica granatum, the sole 
genus with two species only within the family Punicaceae) as follows: 
"To conclude, botanical, historical, and philological data agree in show- 
ing that the modern species is a native of Persia and some adjacent 
countries. Its cultivation began in prehistoric time, and its early 
extension, first toward the west and afterwards into China, has caused 
its naturalization in cases which may give rise to errors as to its true 
origin, for they are frequent, ancient, and enduring." In fact, the 
pomegranate occurs spontaneously in Iran on stony ground, more 
particularly in the mountains of Persian Kurdistan, Baluchistan, and 
Afghanistan. I am in full accord with A. de Candolle's opinion, which, 
as will be seen, is signally corroborated by the investigation that fol- 
lows, and am not in the least disturbed by A. Engler's view 2 that the 
pomegranate occurs wild in Greece and on the islands of the 
Grecian Archipelago, and that, accordingly, it is indigenous in anterior 
Asia and part of the Balkan Peninsula, while its propagation in Italy 
and Spain presumably followed its cultivation in historical times. First, 
as stated also by G. Buschan, 3 these alleged wild trees of Greece are 
not spontaneous, but have reverted from cultivation to a wild state. 4 
Second, be this as it may, all ancient Greek accounts concerning the 
pomegranate relate exclusively to the cultivated, in no case to the 
wild species; and it is a gratuitous speculation of O. Schrader, 5 who 
follows suit with Engler, that the Greek word fro a was originally 
applied to the indigenous wild species, and subsequently transferred 
to the cultivated one. As will be shown hereafter, the Greek term is a 
loan-word. The naturalization of the fruit in the Mediterranean basin 
is, as A. de Candolle justly terms it, an extension of the original 

1 Origin of Cultivated Plants, p. 240. 

2 In Hehn's Kulturpflanzen, p. 246 (8th ed.). 

3 Vorgeschichtliche Botanik, p. 159. 

4 I am unable, however, to share Buschan's view that the wild specimens of Iran 
and north-western India also belong to this class; that area is too extensive to 
allow of so narrow an interpretation. In this case, Buschan is prejudiced in order 
to establish his own hypothesis of an indigenous origin of the tree in Arabia (see 

6 In Hehn's Kulturpflanzen, p. 247. 


The Pomegranate 277 

area; and Hehn is quite right in dating its cultivation on the part of 
the Greeks to a time after the Homeric epoch, and deriving it from Asia 

G. Buschan 1 holds that Europe is out of the question as to the 
indigenous occurrence of the pomegranate, and with regard to Punica 
protopunica, discovered by Balfour on the Island of Socotra, proposes 
Arabia felix as the home of the tree; but he fails to explain the diffusion 
of the tree from this alleged centre. He opposes Loret's conclusions 
with reference to Egypt, where he believes that the tree was naturalized 
from the time of the Eighteenth Dynasty; but he overlooks the prin- 
cipal point made by Loret, namely, that the Egyptian name is a Semitic 
loan-word.' Buschan's theory conflicts with all historical facts, and 
has not been accepted by any one. 

The pomegranate-tree is supposed to be mentioned in the Avesta 
under the name haddnaepata, 3 the wood serving as fuel, and the juice 
being employed in sacrificial libations; but this interpretation is solely 
given by the present ParsI of India and Yezd, and is not certain. The 
fruit, however, is mentioned in Pahlavi literature (above, p. 193). 

There are numerous allusions to the pomegranate of Persia on 
the part of Mohammedan authors and European travellers, and it 
would be of little avail to cite all these testimonies on a subject which 
is perfectly well known. Suffice it to refer to the Fdrs Ndmah* and to 
give the following extract from A. Olearius : 6 — 

"Pomegranate-trees, almond-trees, and fig-trees grow there with- 
out any ordering or cultivation, especially in the Province of Kilan, 
where you have whole forests of them. The wild pomegranates, which 
you find almost every where, especially at Karabag, are sharp or sowrith. 

1 Vorgeschichtliche Botanik, p. 159. 

2 This fact was simultaneously and independently found by an American 
Egyptologist, Ch. E. Moldenke (tfber die in altagyptischen Texten erwahnten 
Baurae, p. 115, doctor dissertation of Strassburg, Leipzig, 1887); so that Loret 
(Flore pharaonique, p. 76) said, "Moldenke est arrive' presque en mSme temps que 
moi, et par des moyens diff events, ce qui donne une entiere certitude a notre d6- 
couverte commune, a la determination du nom egyptien de la grenade." See also 
C. Joret, Plantes dans l'antiquite\ Vol. I, p. 117. Buschan's book appeared in 1895; 
nevertheless he used Loret's work in the first edition of 1887, instead of the second 
of 1892, which is thoroughly revised and enlarged. 

'For instance, Yasna, 62, 9; 68, 1. Cf. also A. V. W. Jackson, Persia Past 
and Present, p. 369. 

4 G. Le Strange, Description of the Province of Fars in Persia, p. 38 (London, 
1912). See also d'Herbelot, Bibliotheque orientale, Vol. Ill, p. 188; and F. Spiegel, 
Eranische Altertumskunde, Vol. I, p. 252. 

8 Voyages of the Ambassadors to the Great Duke of Muscovy, and the Eling 
of Persia (1633-39), P- 2 32 (London, 1669). 

278 Sino-Iranica 

They take out of them the seed, which they call Nardan, wherewith 
they drive a great trade, and the Persians make use of it in their 
sawces, whereto it gives a colour, and a picquant tast, having been 
steep 'd in water, and strain'd through a cloath. Sometimes they boyl 
the juyce of these Pomegranates, and keep it to give a colour to the 
rice, which they serve up at their entertainments, and it gives it withall 
a tast which is not unpleasant. . . . The best pomegranates grow in 
Jescht, and at Caswin, but the biggest, in Karabag." 

Mirza Haidar mentions a kind of pomegranate peculiar to Baluris- 
tan (Kafiristan), sweet, pure, and full-flavored, its seeds being white 
and very transparent. 1 

"Grapes, melons, apples, and pomegranates, all fruits, indeed, are 
good in Samarkand." 2 The pomegranates of Khojand were renowned 
for their excellence. 3 The Emperor Jahangir mentions in his Memoirs 
the sweet pomegranates of Yazd and the subacid ones of Farrah, and 
says of the former that they are celebrated all over the world. 4 J. 
Crawfurd 5 remarks, "The only good pomegranates which, indeed, 
I have ever met with are those brought into upper India by the cara- 
vans from eastern Persia." 

The Yu yan tsa tsu 6 states that the pomegranates of Egypt ty)%W. 
(Wu-se-li, *Mwir-si-li, Mirsir) 7 in the country of the Arabs (Ta-si, 
*Ta-d2ik) weigh up to five and six catties. 

Also in regard to the pomegranate we meet the tradition that its 
introduction into China is due to General Can K'ien. In the same 
manner as in the case of the walnut, this notion looms up only in 
post-Han authors. It is first recorded by Lu Ki ^ SI, who lived under 
the Western Tsin dynasty (a.d. 265-313), in his work Yii ti yiln lu 
H K? § ilF. This text has been handed down in the TsH min yao $u 
of Kia Se-niu of the sixth century. 8 There it is said that Can K'ien, 
while an envoy of the Han in foreign countries for eighteen years, 
obtained Vu-lin %£ $v, this term being identical with nan-H-liu $c 75 
t§?. This tradition is repeated in the Po wu ci 9 of Can Hwa and in the 

1 Elias and Ross, Tarikh-i-Rashidi, p. 386. 

2 A. S. Beveridge, Memoirs of Babur, p. 77. 

8 Ibid., p. 8. They are also extolled by Ye-lu C'u-ts'ai (Bretschneider, Mediae- 
val Researches, Vol. I, p. 19). 

* H. M. Elliot, History of India as told by Its Own Historians, Vol. VI, p. 348 . 

6 History of the Indian Archipelago, Vol. I, p. 433. 
8 81 ft Ch. 10, p. 4 b (ed. of Tsin tai pi Im). 

7 Old Persian Mudraya, Hebrew Mizraim, Syriac Mezroye. 

8 Ch. 4, p. 14 b (new ed., 1896). 

• See above, p. 258. 

The Pomegranate 279 

Tu i U 35 ^ &, written by Li Yu ^ % (or Li Yuan X.) of the Tang 
dynasty. Another formal testimony certifying to the acceptance of 
this creed at that period comes from Fun Yen t^t $1 of the Tang in 
his Fun H wen kien ki ItSfflli 12 , l who states that Can K'ien 
obtained in the Western Countries the seeds of H-liu ?J f@ and alfalfa 
(mu-su), and that at present these are to be found everywhere in 
China. Under the Sung this tradition is repeated by Kao C'eh M *fc. 2 
C'en Hao-tse, in his Hwa kin, 3 published in 1688, states it as a cold- 
blooded fact that the seeds of the pomegranate came from the country 
Nan-si or An-si (Parthia), and that Can K'ien brought them back. 
There is nothing to this effect in Can K'ien's biography, nor is the 
pomegranate mentioned in the Annals of the Han. 4 The exact time of 
its introduction cannot be ascertained, but the tree is on record no earlier 
than the third and fourth centuries a.d. 6 

Li Si-cen ascribes the term nan-H-liu to the Pie lu 3'J &, but he 
cites no text from this ancient work, so that the case is not clear. 6 
The earliest author whom he quotes regarding the subject is T'ao 
Huri-kift (a.d. 452-536), who says, "The pomegranate, particularly as 
regards its blossoms, is charming, hence the people plant the tree in 
large numbers. It is also esteemed, because it comes from abroad. 
There are two varieties, the sweet and the sour one, only the root of 
the latter being used by physicians." According to the TsH min yao $u, 
Ko Huh M $i of the fourth century, in his Pao p'u tse #3 ft -?*, speaks 
of the occurrence of bitter liu ^f $3 on stony mountains. These, indeed, 

1 Ch. 7, p. 1 b (ed. of Ki fu ts'un Su). 

1 jft wu ki yuan ^ $J $6 W> (ed- of Si yin hiian ts'un Su), Ch. 10, p. 34 b. 

* Ch. 3, p. 37, edition of 1783; see above, p. 259. 

* The Can-K'ien legend is repeated without criticism by Bretschneider 
(Bot. Sin., pt. 1, p. 25; pt. 3, No. 280), so that A. de Candolle (Origin of Cultivated 
Plants, p. 238) was led to the erroneous statement that the pomegranate was intro- 
duced into China from Samarkand by Can K'ien, a century and a half before the 
Christian era. The same is asserted by F. P. Smith (Contributions towards the 
Materia Medica of China, p. 176), G. A. Stuart (Chinese Materia Medica, p. 361), 
and Hirth (Toung Pao, Vol. VI, 1895, p. 439). 

8 It is mentioned in the Kin kwei yao lio (Ch. c, p. 27) of the second century a.d., 
"Pomegranates must not be eaten in large quantity, for they injure man's lungs." 
As stated (p. 205), this may be an interpolation in the original text. 

* The Pie lu is not quoted to this effect in the Ceh lei pen ts'ao (Ch. 22, p. 39), 
but the Ci wu min Si t'u k'ao (Ch. 15, p. 102; and 32, p. 36 b) gives two different 
extracts from this work relating to our fruit. In one, its real or alleged medical prop- 
erties are expounded; in the other, different varieties are enumerated, while not a 
word is said about foreign origin. I am convinced that in this form these two texts 
were not contained in the Pie lu. The question is of no consequence, as the work 
itself is lost, and cannot be dated exactly. All that can be said with certainty is that 
it existed prior to the time of T'ao Hun-kin. 

280 Sino-Iranica 

are the particular places where the pomegranate thrives. Su Sun of 
the Sung period states that the pomegranate was originally grown in 
the Western Countries (Si yii^^), and that it now occurs everywhere; 
but neither he nor any other author makes a positive statement as to 
the time and exact place of origin. The Yao sin lun, Pen ts'ao U i, 
and Pen ts x ao yen i l give merely a botanical notice, but nothing of his- 
torical interest. 

The pomegranate (U-liu) is mentioned in the "Poem on the Capital 
of Wu" ^1 %$ i® by Tso Se & JB», who lived in the third century under 
the Wu dynasty (a.d. 222-280). P'an Yo 1 S, a poet of the fourth 
century a.d., says, "Pomegranates are the most singular trees of the 
empire and famous fruits of the Nine Provinces. 2 A thousand seed- 
cases are enclosed by the same membrane, and what looks like a single 
seed in fact is ten." 

The Tsin Lun nan k'i kii lu If B§ :2c %. M 3: ("Annotations on 
the Conditions of the period Lun-nan [a.d. 397-402] of the Tsin Dy- 
nasty") contains the following note: 3 "The pomegranates (nan Si 
liu) of the district Lin-yuan E$& ^7C in Wu-lin l£ §^ 4 are as large as cups; 
they are not sour to the taste. Each branch bears six fruits." 

Lu Hui $&$M of the Tsin dynasty, in his Ye lun ki W$> ^tffi, 5 states 
that in the park of Si Hu 15 ^ there were pomegranates with seeds as 
large as cups, and they were not sour. Si Hu or Si Ki-lun ^ ^ ft ruled 
from a.d. 335 to 349, under the appellation T'ai Tsu ^C IB of the Hou 
Cao dynasty, as "regent celestial king" (kii-te Vien wan), and shifted 
the capital to Ye ^, the present district of Lin-can W> W, in the pre- 
fecture of Cah-te ^ IS in Ho-nan. 6 

The pomegranate is mentioned in the Ku kin lu ~& ^t 1&, 7 written 
by Ts'ui Pao $L 15 during the middle of the fourth century, with 
reference to the pumelo tti" (Citrus grandis), the fruit of which is com- 
pared in shape with the pomegranate. The TsH min yao lu (I.e.) gives 
rules for the planting of pomegranates. 

1 Ch. 18, p. 7 (ed. of Lu Sin-yuan); the other texts see in Cen lei pen ts'ao, I. c. 

2 JL 'M , the ancient division of China under the Emperor Yu. 

8 T'ai p'ih yil Ian, Ch. 970, p. 4 b. Regarding the department of records styled 
k'i kii lu, see The Diamond, p. 35. In the Yuan kien lei han (Ch. 402, p. 2) the 
same text is credited to the Sun Su. 

4 In Hu-nan Province. 

6 Ed. of Wu yin tien, p. 12. 

6 Regarding his history, see L. Wieger, Textes historiques, pp. 1095-1 100. 
Bretschneider's (Bot. Sin., pt. 1, p. 211) note, that, besides the Ye Sun ki of Lu 
Hui, there isjanother work of the same name by Si Hu, is erroneous; Si Hu is simply 
the "hero" of the Ye lun ki. 

7 Ch. C, p. 1 (ed. of Han Wei ts'un Su or Ki fu ts'un Su). Cf. also below, p. 283. 

The Pomegranate 281 

The Annals of the Liu Sung Dynasty, a.d. 420-477 (SunSu), contain 
the following account: "At the close of the period Yiian-kia jt % 
(a.d. 424-453), when T'ai Wu (a.d. 424-452) ^C i£ of the Wei dynasty 
conquered the city Ku Wi m* he issued orders to search for sugar- 
cane and pomegranates (nan H liu). Can C'ah 3fc !§ said that pome- 
granates (H-liu) come from Ye." This is the same locality as mentioned 

The Stan kwo ki H H HE 2 reports that in the district of Lun-kan 
ft [33 M 3 there are good pomegranates (H liu). These various examples 
illustrate that in the beginning the tree was considered as peculiar to 
certain localities, and that accordingly a gradual dissemination must 
have taken place. Apparently no ancient Chinese author is informed 
as to the locality from which the tree originally came, nor as to the how 
and when of the transplantation. 

The Kwan U & ;£, written by Kwo Yi-kun M ^ ^ prior to a.d. 
527, as quoted in the Ts'i min yao $u, discriminates between two varie- 
ties of pomegranate (nan H liu), a sweet and a sour one, in the same 
manner as T'ao Hun-kin. 4 This distinction is already made by Theo- 
phrastus. 6 As stated above, there was also a bitter variety. 6 

It is likewise a fact of great interest that we have an isolated instance 
of the occurrence of a pomegranate-tree that reverted to the wild state. 
The Lii San ki % \h Ifi 7 contains this notice: "On the summit of the 
Hian-lu fun § m % Censer-Top ') there is a huge rock on which 
several people can sit. There grows a wild pomegranate (San H-liu 
tf] 3? Jf?) drooping from the rock. In the third month it produces blos- 
soms. In color these resemble the [cultivated] pomegranate, but they 

1 Modern Cen-tin fu in Ci-li Province. 

2 Thus in T'ai p'in yii Ian, Ch. 970, p. 5 b; the Ts'i min yao Su (Ch. 4, p. 14) 
ascribes the same text to the Kin k'ou ki 3^ P ft}. 

3 At present the district which forms the prefectural city of Sun-te in Ci-li 

4 Above, p. 279. 

6 Historia plantarum, II. n, 7. 

6 Pliny (XIII, 113) distinguishes five varieties, — dulcia, acria, mixta, acida, 

7 T'ai p'in yii Ian, Ch. 970, p. 5. The Lu Mountain is situated in Kiah-si Prov- 
ince, twenty-five li south of Kiu-kian. A work under the title Lii San ki was written 
by C'en Lin-ku ^ ^ ^L in the eleventh century (Wylie, Notes on Chinese Liter- 
ature, p. 55); but, as the T'ai p'in yii Ian was published in a.d. 983, the question here 
must be of an older work of the same title. In fact, there is a Lii San ki by Kin Si 
Jr 5^ of the Hou Cou dynasty; and the Yuan kien lei han (Ch. 402, p. 2) ascribes 
the same text to the Cou Kin Si Lii San ki. The John Crerar Library of Chicago 
(No. 156) possesses a Lii San siao li in 24 chapters, written by Ts'ai Yin ^| ^ and 
published in 1824. 

282 Sino-Iranica 

are smaller and pale red. When they open, they display a purple calyx 
of bright and attractive hues." A poem of Li Te-yu ^ £§ $r (787-849) 
opens with the words, "In front of the hut where I live there is a wild 
pomegranate." 1 

Fa Hien ££ H, the celebrated Buddhist traveller, tells in his Fu kwo 
ki ffi H 12 ("Memoirs of Buddhist Kingdoms"), written about a.d. 
420, that, while travelling on the upper Indus, the flora differed from 
that of the land of Han, excepting only the bamboo, pomegranate, and 
sugar-cane.' This passage shows that Fa Hien was familiar with that 
tree in China. Huan Tsan observed in the seventh century that pome- 
granates were grown everywhere in India. 3 Soleiman (or whoever may 
be the author of this text), writing in a.d. 851, emphasizes the abun- 
dance of the fruit in India. 4 Ibn Batuta says that the pomegranates of 
India bear fruit twice a year, and emphasizes their fertility on the 
Maldive Islands. 6 Seedless pomegranates came to the household of the 
Emperor Akbar from Kabul. 6 

The pomegranate occurred in Fu-nan (Camboja), according to the 
Nan TsH i« or History of the Southern Ts'i (a.d. 479-501), compiled 
by Siao Tse-hien in the beginning of the sixth century. 7 It is mentioned 
again by Cou Ta-kwanof the Yuan dynasty, in his book on the "Customs 
of Camboja." 8 In Han-cou, large and white pomegranates were styled 
yil liu 3£ fH? ("jade" liu), while the red ones were regarded as inferior or 
of second quality. 8 

The following ancient terms for the pomegranate, accordingly, are 
on record: — 

(1) Wl # Vu-lin, *du-lim. Aside from the Po wu li, this term is 
used by the Emperor Yuan of the Liang dynasty in a eulogy of the 
fruit. 10 Hirth 11 identified this word with an alleged Indian daritn; and, 
according to him, Can K'ien must have brought the Indian name to 

' Li wei kun pie tsi, Ch. 2, p. 8 {Ki fu ts'un Su, t'ao 10). 
2 Cf. J. Legge, A Record of Buddhistic Kingdoms, p. 24. 

* Ta T'an si yil ki, Ch. 2, p. 8 b (S. Beal, Buddhist Records of the Western 
World, Vol. I, p. 88). 

4 M. Reinaud, Relation des voyages, Vol. I, p. 57. 

6 Defremery and Sanguinetti, Voyages d'Ibn Batoutah, Vol. Ill, p. 129. 

* H. Blochmann, Ain I Akbari, Vol. I, p. 65. 

7 Pelliot, Le Fou-nan, Bull, de VEcole francaise, Vol. Ill, p. 262. 

8 Pelliot, ibid., Vol. II, p. 168. 

* Mori lian /« ^ ^ H by Wu Tse-mu % g #C of the Sung (Ch. 18, p. 5 b; 
ed. of Ci pu tsu lai ts'un Su). 

10 Yuan kien lei han, Ch. 402, p. 3 b. Further, in the lost Hu pen ts'ao, as follows 
from a quotation in a note to the Pei hu lu (Ch. 3, p. 12). 

11 T'oung Pao, Vol. VI, 1895, p. 439. 

The Pomegranate 283 

China. How this would have been possible, is not explained by him. 
The Sanskrit term for the pomegranate (and this is evidently what 
Hirth hinted at) is dddima or ddlima, also dddimva, which has passed 
into Malayan as delima. 1 It is obvious that the Chinese transcription 
bears some relation to this word; but it is equally obvious that the 
Chinese form cannot be fully explained from it, as it leads only to 
*du-lim, not, however, to dalim. There are two possibilities: the Chinese 
transcription might be based either on an Indian vernacular or 
Apabhramca form of a type like *dulim, *ducjim, 2 or on a word of the 
same form belonging to some Iranian dialect. The difficulty of the 
problem is enhanced by the fact that no ancient Iranian word for the 
fruit is known to us. 3 It appears certain, however, that no Sanskrit 
word is intended in the Chinese transcription, otherwise we should 
meet the latter in the Sanskrit-Chinese glossaries. The fact remains 
that these, above all the Fan yi mih yi tsi, do not contain the word 
Vu-lin; and, as far as I know, Chinese Buddhist literature offers no 
allusion to the pomegranate. Nor do the Chinese say, as is usually 
stated by them in such cases, that the word is of Sanskrit origin; the 
only positive information given is that it came along with General 
Can K'ien, which is to say that the Chinese were under the im- 
pression that it hailed from some of the Iranian regions visited by him. 
*Dulim, dulima, or *durim, durima, accordingly, must have been a 
designation of the pomegranate in some Iranian language. 

(2) :$" ?* tan-So, *dan-zak, dan-yak, dan-n'iak. This word appears 
in the Ku kin lu K and in the Yu yah tsa tsu. 6 Apparently it represents a 
transcription, but it is not stated from which language it is derived. In 
my estimation, the foundation is an Iranian word still unknown to us, 
but congeners of which we glean from Persian ddnak ("small grain"), 

1 J. Crawfurd (History of the Indian Archipelago, Vol. I, p. 433) derives this 
word from the Malayan numeral five, with reference to the five cells into which the 
fruit is divided. This, of course, is a mere popular etymology. There is no doubt 
that the fruit was introduced into the Archipelago from India; it occurs there only 
cultivated, and is of inferior quality. On the Philippines it was only introduced 
by the Spaniards (A. de Morga, Philippine Islands, p. 275, ed. of Hakluyt Society). 

2 The vernacular forms known to me have the vowel a; for instance, Hindustani 
darim, Bengali dalim, dalim or darim; Newari, dhdde. The modern Indo-Aryan 
languages have also adopted the Persian word anar. 

* In my opinion, the Sanskrit word is an Iranian loan-word, as is also Sanskrit 
karaka, given as a synonyme for the pomegranate in the Amarakosa. The earliest 
mention of dddima occurs in the Bower Manuscript; the word is absent in Vedic 

4 At least it is thus stated in cyclopaedias; but the editions of the work, as 
reprinted in the Han Wei ts'un Su and Kifu ts'un Su, do not contain this term. 

» Ch. 18, p. 3 b (ed. of Pai hai). 

284 Sino-Iranica 

dana ("grain, berry, stone of a fruit, seed of grain or fruit"), ddngu 
("kind of grain"), Sina danu ("pomegranate"); 1 Sanskrit dhanika, 
dhanydka, or dhaniyaka ("coriander"; properly "grains"). The no- 
tion conveyed by this series is the same as that underlying Latin 
granatum, from granum ("grain"); cf. Anglo-Saxon corn&ppel and 
English pomegranate ("apple made up of grains"). 

(3) 5: ^B fa? nan H Uu or ^ ^ H Uu. This transcription is generally 
taken in the sense "the plant Uu of the countries Nan and Si, or of the 
country Nan-Si." This view is expressed in the Po wu U, which, as 
stated, also refers to the Can-K'ien legend, and to the term t'u-lin, 
and continues that this was the seed of the Uu of the countries Nan 
and Si; hence, on the return of Can K'ien to China, the name nan-H-liu 
was adopted. 2 Bretschneider intimates that Nan and Si were little 
realms dependent on K'an at the time of the Han. Under the T'ang, 
the name Nan referred to Bukhara, and Si to TaSkend; but it is hardly 
credible that these two geographical names (one does not see for what 
reason) should have been combined into one, in order to designate 
the place of provenience of the pomegranate. It is preferable to assume 
that j£: 3? nan H, *an-sek, an-sak, ar-sak, represents a single name 
and answers to Arsak, the name of the Parthian dynasty, being on a 
par with j£c & nan-si, *Ar-sik, and :£ US nan-si, *Ar-sai. In fact, 
^ ^ is the best possible of these transcriptions. We should expect, 
of course, to receive from the Chinese a specific and interesting story as 
to how and when this curious name, which is unique in their botanical 
nomenclature, was transmitted; 3 but nothing of the kind appears to 
be on record, or the record, if it existed, seems to have been lost. It 
is manifest that also the plant-name Uu (*riu, r'u) presents the tran- 
scription of an Iranian word, and that the name in its entirety was 
adopted by the Chinese from an Iranian community outside of Parthia, 
which had received the tree or shrub from a Parthian region, and there- 
fore styled it "Parthian pomegranate." It is not likely that the tree 
was transplanted to China directly from Parthia; we have to assume 
rather that the transplantation was a gradual process, in which the 

1 W. Leitner, Races and Languages of Dardistan, p. 17. 

2 It is not correct, as asserted by Bretschneider (Chinese Recorder, 1871, 
p. 222), to say that this definition emanates from Li Si-c"en, who, in fact, quotes 
only the Po wu It, and presents no definition of his own except that the word Uu 
means ^ liu ("goitre"); this, of course, is not to be taken seriously. In Jehol, a 
variety of pomegranate is styled hai $£ Uu (O. Franke, Beschreibung des Jehol- 
Gebietes, p. 75); this means literally, "Uu from the sea," and signifies as much as 
"foreign Uu." 

3 Cf . nan-si Man $£ J& ^ ("Parthian incense") as designation for styrax 
benzoin (p. 464). 

The Pomegranate 285 

Iranian colonies outside of Iran proper, those of Sogdiana and Turkis- 
tan, played a prominent part. We know the Sogdian word for the 
pomegranate, which is written n'r'kh, and the reading of which has 
been reconstructed by R. Gauthiot 1 in the form *narak(a), developed 
from *anar-aka. This we meet again in Persian andr, which was adopted 
in the same form by the Mongols, while the Uigur had it as nara. At 
all events, however, it becomes necessary to restore, on the basis of the 
Chinese transcription, an ancient *riu, *ru, of some Iranian dialect. 
This lost Iranian word, in my opinion, presents also the foundation of 
Greek ?>6a or /Joid, — the origin of which has been hitherto unexplained or 
incorrectly explained, 2 — and the Semitic names, Hebrew ritnmon, 
Arabic rumman, Amharic riiman, Syriac rumono, Aramaic rummana, 
from which Egyptian arhmani or anhmdni (Coptic ertnan or hertnan) / 

is derived. 3 

(4) ^B %H lo-liu, *zak (yak, n'iak)-liu (riu). This hybrid compound, 
formed of elements contained in 2 and 3, is found in the dictionary 
Kwan ya M 5ft, written by Can Yi 181 M about a.d. 265. 4 It is also 
employed by the poet P'an Yo of the fourth century, mentioned above. 5 
Eventually also this transcription might ultimately be traced to an 
Iranian prototype. Japanese zakuro is based on this Chinese form. 6 

While the direct historical evidence is lacking, the Chinese names of 
the tree point clearly to Iranian languages. Moreover, the tree itself, 
is looked upon by the Chinese as a foreign product, and its first intro-j 
duction into China appears to have taken place in the latter part of; 
the third century a.d. 

In my opinion, the pomegranate-tree was transplanted to India, 

1 Essai sur le vocalisme du sogdien, p. 49. Cf. also Armenian nrneni for the 
tree and nurn for the fruit. 

2 The etymologies of the Greek word enumerated by Schrader (in Hehn, 
Kulturpflanzen, p. 247) are so inane and far-fetched that they do not merit dis- 
cussion. It is not necessary, of course, to hold that an immediate transmission of 
the Persian word took place, but we must look to a gradual propagation and to 
missing links by way of Asia Minor. According to W. Muss-Arnolt (Transactions 
Am. Phil. Assoc, Vol. XXIII, 1892, p. no), the Cyprian form fivdia forbids all 
connection with the Hebrew. It is not proved, however, that this dialectic word 
has any connection with /J6a; it may very well be an independent local development. 

3 V. Loret, Flore pharaonique, p. 76. Portuguese roma, romeira, from the 
Arabic; Anglo-Saxon read-appel. 

4 This is the date given by Waiters (Essays on the Chinese Language, p. 38). 
Bretschneider (Bot. Sin., pt. I, p. 164) fixes the date at about 227-240. 

B T'an lei han, Ch. 183, p. 9. 

6 Written also ffi ;R§. E. Kaempfer (Amoenitates exoticae, p. 800) already 
mentions this term as dsjakurjo, vulgo sakuro, with the remark, "Rara est hoc 
coelo et fructu ingrato." 

286 Sino-Iranica 

likewise from Iranian regions, presumably in the first centuries of 
our era. The tree is not mentioned in Vedic, Pali, or early Sanskrit 
literature; and the word dalima, dddima, etc., is traceable to Iranian 
*dulim(a), which we have to reconstruct on the basis of the Chinese 
transcription. The Tibetans appear to have received the tree from 
Nepal, as shown by their ancient term bal-poi seu-Un ("seu tree of 
Nepal")- 1 From India the fruit spread to the Malayan Archipelago 
and Camboja. Both Cam dalim and Khmer tatitn 2 are based on the 
Sanskrit word. The variety of pomegranate in the kingdom of Nan-cao 
in Yun-nan, with a skin as thin as paper, indicated in the Yu yan tsa 
tsu, 3 may also have come from India. J. Anderson 4 mentions pome- 
granates as products of Yun-nan. 

Pomegranate-wine was known throughout the anterior Orient at 
an early date. It is pointed out under the name asls in Cant. VIII, 2 
(Vulgata: musturn) and in the Egyptian texts under the name ledek-it} 
Dioscorides 6 speaks of pomegranate- wine (potrr/s olvos). Ye-lu C'u- 
ts'ai, in his Siyulu (account of his journey to Persia., 1219-24), speak- 
ing of the pomegranates of Khojand, which are "as large as two fists 
and of a sour-sweet taste," says that the juice of three or five fruits is 
pressed out into a vessel and makes an excellent beverage. 7 In the 
country Tun-sun HI 21 (Tenasserim) there is a wine-tree resembling 
the pomegranate; the juice of its flowers is gathered and placed in jars, 
whereupon after several days it turns into good wine. 8 The inhabitants 
of Hai-nan made use of pomegranate-flowers in fermenting their wine. 9 
I have not found any references to pomegranate-wine prepared by the 
Chinese, nor is it known to me that they actually make such wine. 

It is known that the pomegranate, because of its exuberant seeds, 
is regarded in China as an emblem alluding to numerous progeny; it 
has become an anti-race-suicide symbol. The oldest intimation of this 
symbolism looms up in the Pei H At $., where it is told that two pome- 
granates were presented to King Nan-te ;2c M of Ts'i ^ on the occasion 

1 This matter has been discussed by me in T'oung Pao, 1916, pp. 408-410. In 
Lo-lo we have sa-bu-se in the A-hi dialect and se-bu-se in Nyi. Sa or se means "grain " 
(corresponding to Tibetan sa in sa-bon, "seed"). The last element se signifies 
"tree." The fruit is se-bu-ma (ma, "fruit"). 

• Aymonier and Cabaton, Dictionnaire cam-francais, p. 220. 

• Ch. 18, p. 3 b. 

4 Report on the Expedition to Western Yunan, p. 93 (Calcutta, 1871). 

• V. Loret, Flore pharaonique, pp. 77, 78. 
' v, 34- 

7 Bretschneider, Mediaeval Researches, Vol. I, p. 19. 

8 Liafi Su, Ch. 54, p. 3. 

• Hirth, Chau Ju-kua, p. 177. 


The Pomegranate 287 

of his marriage to the daughter of Li Tsu-sou ^ IB. *&. The latter 
explained that the pomegranate encloses many seeds, and implies the 
wish for many sons and grandsons. Thus the fruit is still a favorite 
marriage gift or plays a r61e in the marriage feast. 1 The same is the 
case in modern Greece. Among the Arabs, the bride, when dismounting 
before the tent of the bridegroom, receives a pomegranate, which she 
smashes on the threshold, and then flings the seeds into the interior of 
the tent. 1 The Arabs would have a man like the pomegranate, — bitter- 
sweet, mild and affectionate with his friends in security, but tempered 
with a just anger if the time call him to be a defender in his own or in 
his neighbor's cause. 3 

1 See, for instance, H. Dor£, Recherches sur les superstitions en Chine, pt. I 
Vol. II, p. 479. 

* A. Musil, Arabia Petraea, Vol. Ill, p. 191. 

■ C. M. Doughty, Travels in Arabia Deserta, Vol. I, p. 564. 


6. In A. de Candolle's book 1 we read, "Chinese works seem to 
show that sesame was not introduced into China before the Christian 
era. The first certain mention of it occurs in a book of the fifth or sixth 
century, entitled TsH min yao £w. Before this there is confusion between 
the name of this plant and that of flax, of which the seed also yields an 
oil, and which is not very ancient in China." Bretschneider is cited as 
the source for this information. It was first stated by the latter that, 
according to the Pen ts'ao, hu ma ffl M (Sesamum orientate) was brought 
by Can K'ien from Ta-yiian. 2 In his "Botanicon Sinicum" 3 he asserts 
positively that hu ma, or foreign hemp, is a plant introduced from west- 
ern Asia in the second century B.C. 4 The same dogma is propounded 
by Stuart. 6 

All that there is to this theory amounts to this. T'ao Huh-kin 
(a.d. 451-536) is credited in the Pen ts'ao kan mu* with the statement 
that "huma ffli ffc ('hemp of the Hu') originally grew in Ta-yiian 
(Fergana) >|C £fe ^c $J, 7 and that it henCe received the name hu ma 
('Iranian hemp ')•" He makes no reference to Can K'ien or to the time 
when the introduction must have taken place; and to every one 
familiar with Chinese records the passage must evoke suspicion through 
its lack of precision and chronological and other circumstantial evi- 
dence. The records regarding Ta-yiian do not mention hu ma, nor 
does this term ever occur in the Annals. Now, T'ao Hun-kih was a 
Taoist adept, a drug-hunter and alchemist, an immortality fiend; he 
never crossed the boundaries of his country, and certainly had no 
special information concerning Ta-yiian. He simply drew on his 
imagination by arguing, that, because mu-su (alfalfa) and grape sprang 

1 Origin of Cultivated Plants, p. 420. 

2 Chinese Recorder, 1871, p. 222; adopted by Hirth, T'oung Pao, Vol. VI, 1895, 
p. 439, and maintained again in Journal Am. Or. Soc, 1917, p. 92. 

3 Pt. II, p. 206. 

4 Ibid., p. 204, he says, however, that the Pen ts'ao does not speak of flax, and 
that its introduction must be of more recent date. This conflicts with his statement 

5 Chinese Materia Medica, p. 404. 

6 Ch. 22, p. 1. Likewise in the earlier Gen lei pen ts'ao, Ch. 24, p. 1 b. 

7 This tradition is reproduced without any reference in the Pen ts'ao yen i of 
1 1 16 (Ch. 20, p. 1, ed. of Lu Sin-yuan). 


Sesame and Flax 289 

from Ta-yuan (that is, a Hu country), hu ma also, being a Hu plant, 
must likewise have emanated from that quarter. Such vagaries 
cannot be accepted as history. All that can be inferred from the passage 
in question is that T'ao Huh-kin may have been familiar with hu ma. 
Li Si-cen, quoting the Mon k% pi Van ^ H ¥ 1^ by Sen Kwa tt tS 1 
of the eleventh century, says, "In times of old there was in China only 
'great hemp' ta ma ^Ciftt (Cannabis sativa) growing in abundance. 
The envoy of the Han, Can K'ien, was the first to obtain the seeds of 
oil-hemp Y& M 2 from Ta-yuan; hence the name hu ma in distinction 
from the Chinese species ta ma." The Can -K'ien tradition is further 
voiced in the T'un U of Cen Tsiao (1108-62) of the Sung. 3 The T'ai 
p'in yii /aw, 4 published in a.d. 983, quotes a Pen ts'ao kin of unknown 
date as saying that Can K'ien obtained from abroad hu ma and hu tou. b 
This legend, accordingly, appears to have arisen under the Sung (a.d. 
960-1278); that is, over a millennium after Can K'ien's lifetime. And 
then there are thinking scholars who would make us accept such stuff 
as the real history of the Han dynasty! 

In the T'ang period this legend was wholly unknown: the T'ah Pen 
ts'ao does not allude to any introduction of hu ma, nor does this work 
speak of Can K'ien in this connection. 

A serious book like the T'u kin pen ts'ao of Su Sun, which for the 
first time has also introduced the name yu ma ("oil hemp"), says only 
that the plant originally grew in the territory of the Hu, that in appear- 
ance it is like hemp, and that hence it receives the name hu ma. 

Unfortunately it is only too true that the Chinese confound Sesamum 
indicum (family Pedaliaceae) and Linum usitatissimum (family Linaceae) 
in the single term hu ma ("Iranian hemp"); the only apparent reason 
for this is the fact that the seeds of both plants yield an oil which is put 
to the same medicinal use. The two are totally different plants, nor 
do they have any relation to hemp. Philologically, the case is somewhat 
analogous to that of hu tou (p. 305). It is most probable that the two 
are but naturalized in China and introduced from Iranian regions, for 
both plants are typically ancient West-Asiatic cultivations. The alleged 
wild sesame of China 6 is doubtless an escape from cultivation. 

1 This is the author wrongly called "Ch'en Ts'ung-chung " by Bretschneider 
(Bot. Sin., pt. II, p. 377). Ts'un-aun # 4* is his hao. 
1 A synonyme of hu ma. 
8 Ch. 75, p. 33. 

4 Ch. 841, p. 6 b. 

5 See below, p. 305. 

6 Forbes and Hemsley, Journal Linnean Soc, Vol. XXVI, p. 236. 

290 Sino-Iranica 

Herodotus 1 emphasizes that the only oil used by the Babylonians 
is made from sesame. Sesame is also mentioned among their products 
by the Babylonian priest Berosus (fourth century B.C.). 2 

Aelius Gallus, a member of the Equestrian order, carried the Roman 
arms into Arabia, and brought back from his expedition the report that 
the Nomades (nomads) live on milk and the flesh of wild animals, and 
that the other peoples, like the Indians, express a wine from palms and 
oil from sesame. 3 According to Pliny, sesame comes from India, where 
they make an oil from it, the color of the seeds being white. 4 Both the 
seeds and the oil were largely employed in Roman pharmacology. 5 
Megasthenes 6 mentions the cultivation of sesame in India. It likewise 
occurs in the Atharva Veda and in the Institutes of Manu (Sanskrit 
tila). 7 A. de Candolle's view 8 that it was introduced into India from 
the Sunda Isles in prehistoric times, is untenable. This theory is based 
on a purely linguistic argument: "Rumphius gives three names for 
the sesame in these islands, very different one from the other, and from 
the Sanskrit word, which supports the theory of a more ancient existence 
in the archipelago than on the continent." This alleged evidence proves 
nothing whatever for the history of the plant, but is merely a fact of 
language. 9 There can now be no doubt that from a botanical viewpoint 
the home of the genus is in tropical Africa, where twelve species occur, 
while there are only two in India. 10 

In the Fan yi min yi tsi, 11 a Sanskrit synonyme of "sesame" is given as 
M $1 @ & ife a-Vi-mu-to-k x ie f *a-di-muk-ta-g'a, i.e., Sanskrit adhi- 
muktaka, which is identified with kil-hn (see below) and hu-ma. An 
old gloss explains the term as "the foreign flower of pious thoughtful- 
ness" (§an se i hwa W M 3^1 ^), an example of which is the lighting of 
a lamp fed with the oil of three flowers (sandal, soma, and campaka 
[Michelia champaca]) and the placing of this lamp on the altar of the 

1 1. 193- 

2 Muller, Fragmenta historiae graecae, Vol. II, p. 496. Regarding Egypt, 
see V. Loret, Flore pharaonique, p. 57. 

8 Pliny, vi, 28, §161. 

4 Sesama ab Indis venit. Ex ea et oleum faciunt; colos eius candidus (xvm, 
22, §96). 

6 Pliny, xxii, 64, §132. 

6 Strabo, XV. I, 13. 

7 Joret, Plantes dans l'antiquite\ Vol. II, p. 269. 

8 Origin of Cultivated Plants, p. 422. 

9 The Malayan languages possess a common name for Sesamum indicum: 
Javanese and Malayan Una, Batak Una, Cam lon'6 or land; Khmer lono. 

10 A. Engler, Pflanzenfamilien, Vol. IV, pt. 3 b, p. 262. 

11 Ch. 8, p. 6 (see above, p. 254). 

Sesame and Flax 291 

Triratna. 1 From the application of adhimuktaka it becomes self-evident 
also that sesame-oil must be included in this series. The frequent 
mention of this oil for sacred lamps is familiar to all readers of the 
Buddhist Jataka. The above Sanskrit-Chinese Dictionary adds the 
following comment: "This plant is in appearance like the 'great hemp' 
(Cannabis sativa). It has red flowers and green leaves. Its seeds can 
be made into oil; also they yield an aromatic. According to the Tsun 
kin yin nie lun ^ Hi §1 $£ p&, sesame (kii-Sen) is originally charcoal, 
and, while for a long time buried in the soil, will change into sesame. 
In the western countries (India) it is customary in anointing the body 
with fragrant oil to use first aromatic flowers and then to take sesame- 
seeds. These are gathered and soaked till thoroughly bright ; afterwards 
they proceed to press the oil out of the sesame, which henceforth be- 
comes fragrant." 

Of greater importance for our purpose is the antiquity of sesame in 
Iran. According to Herodotus 2 , it was cultivated by the Chorasmians, 
Hyrcanians, Parthians, Sarangians, and Thamanaeans. In Persia 
sesame-oil was known at least from the time of the first Achaemenides. 3 
G. Watt 4 even looks to Persia and Central Asia as the home of the 
species; he suggests that it was probably first cultivated somewhere 
between the Euphrates valley and Bukhara south to Afghanistan and 
upper India, and was very likely diffused into India proper and the 
Archipelago, before it found its way to Egypt and Europe. 

Sesamum indicum (var. subindivisum Dl.) is cultivated in Russian 
Turkistan and occupies there the first place among the oil-producing 
plants. It thrives in the warmest parts of the valley of Fergana, and 
does not go beyond an elevation of two thousand five hundred feet. 
It is chiefly cultivated in the districts of Namanga and Andijan, though 
not in large quantity. 6 Its Persian name is kunjut. 

While there is no doubt that this species was introduced into China 
from Iranian regions, the time as to when this introduction took place 
remains obscure. First, there is no historical and dependable record 
of this event; second, the confusion brought about by the Chinese in 
treating this subject is almost hopeless. Take the earliest notice of 
hu ma cited by the Pen ts'ao and occurring in the Pie lu: "Hu ma is 
also called kit-Sen E )$. It grows on the rivers and in the marshes of 

1 Cf. Eitel, Handbook of Chinese Buddhism, p. 4. 
1 in, 117. 

8 Joret, op. cil., Vol. II, p. 71. Sesame is mentioned in Pahlavi literature 
(above, p. 193). 

4 Gingelly or Sesame Oil, p. 11 (Handbooks of Commercial Products, No. 21). 
1 S. Korzinski, Vegetation of Turkistan (in Russian), p. 50. 

292 Sino-Iranica 

San-tan Ji $& (south-eastern portion of San-si), and is gathered in the 
autumn. What is called ts'in San # jj| are the sprouts of the feu-fen. 
They grow in the river-valleys of Cuh-yuan 4* 0. (Ho-nan)." Nothing 
is said here about a foreign introduction or a cultivation; on the con- 
trary, the question evidently is of an indigenous wild swamp-plant, 
possibly Mulgedium sibiriacum. 1 Both Sesamum and Linurn are thor- 
oughly out of the question, for they grow in dry loam, and sesame espe- 
cially in sandy soil. Thus suspicion is ripe that the terms hu ma and 
kU-$en originally applied to an autochthonous plant of San-si and 
Ho-nan, and that hu ma in this case moves on the same line as the term 
hu leh in the Li sao (p. 195). This suspicion is increased by the fact 
that hu ma occurs in a passage ascribed to Hwai-nan-tse, who died in 
122 b.c, and cited in the T x ai p'in yil Ian. 2 Moreover, the Wu H (or 
p % u) pen ts'ao, written in the first half of the third century by Wu P'u 
zik ^, in describing hu ma, alludes to the mythical Emperor Sen-nun 
and to Lei kuh Is &, a sage employed by the Emperor Hwan in his 
efforts to perfect the art of healing. 

The meaning of kit-Sen is "the great superior one." The later authors 
regard the term as a variety of Sesamum, but give varying definitions 
of it: thus, T'ao Huh-kih states that the kind with a square stem is 
called kit-Sen (possibly Mulgedium), that with a round stem hu ma. 
Su Kuh of the T'ang says that the plant with capsules (kio $} ) of eight 
ridges or angles {pa len A Wt) is called kit-ten; that with quadrangular 
capsules, hu ma. The latter definition would refer to Sesamum indicum, 
the capsule of which is oblong quadrangular, two-valved and two-celled, 
each cell containing numerous oily seeds. 

Moh Sen Jlil ffe, in his Si Xiao pen Vsao (written in the second half 
of the seventh century), observes that "the plants cultivated in fertile 
soil produce octangular capsules, while those planted in mountainous 
fields have the capsules quadrangular, the distinction arising from the 
difference of soil conditions, whereas the virtues of the two varieties are 
identical. Again, Lei Hiao H!" Q% of the fifth century asserts that 
kit-Sen is genuine, when it has seven ridges or angles, a red color, and 
a sour taste, but that it is erroneous to style hu ma the octangular 
capsules with two pointed ends, black in color, and furnishing a black oil. 
There is no doubt that in these varying descriptions entirely different 
plants are visualized. Kao C'eh of the Sung, in his Si wu ki yiian, 3 

1 Stuart, Chinese Materia Medica, p. 269. This identification, however, is 

1 Ch. 989, p. 6 b. 

3 Ch. 10, p. 29 b (see above, p. 279). 

Sesame and Flax 293 

admits that it is unknown what the hu ma spoken of in the Pen-ts % ao 
literature really is. 

I have also prepared a translation of Li §i-cen's text on the subject, 
which Bretschneider refrained from translating; but, as there are several 
difficult .botanical points which I am unable to elucidate, I prefer to 
leave this subject to a competent botanist. In substance Li Si-cen 
understands by hu ma the sesame, as follows from his use of the modern 
term U ma la M. He says that there are two crops, an early and a late 
one, 1 with black, white, or red seeds; but how he can state that the 
stems are all square is unintelligible. The criticism of the statements 
of his predecessors occupies much space, but I do not see that it enlight- 
ens us much. The best way out of this difficulty seems to me Stuart's 
suggestion that the Chinese account confounds Sesamum, Linum, 
and Mulgedium. The Japanese naturalist Ono Ranzan 2 is of the same 
opinion. He says that there is no variety of sesame with red seed, as 
asserted by Li Si-cen (save that the black seeds of sesame are reddish 
in the immature stage), and infers that this is a species of Linum which 
always produces red seeds exclusively. Ono also states that there is a 
close correlation between the color of the seeds and the angles of the 
capsules: a white variety will always produce two or four-angled cap- 
sules, while hexangular and octangular capsules invariably contain only 
black seeds. Whether or in how far this is correct I do not know. The 
confusion of Sesamum and Linum arose from the common name hu ma, 
but unfortunately proves that the Chinese botanists, or rather pharma- 
cists, were bookworms to a much higher degree than observers; for it 
is almost beyond comprehension how such radically distinct plants 
can be confounded by any one who has even once seen them. In view 
of this disconsolate situation, the historian can only beg to be excused. 

7. It is a point of great culture-historical interest that the Chinese 
have never utilized the ,flax-fibre in the manufacture of textiles, but j 
that hemp has always occupied this place from the time of their j 
earliest antiquity. 8 This is one of the points of fundamental diversity 
between East-Asiatic and Mediterranean civilizations, — there hemp, 
and here flax, as material for clothing. There are, further, two important \ 
facts to be considered in this connection, — first, that the Aryans 

1 In S. Couling's Encyclopaedia Sinica (p. 504) it is stated that in China there is 
only one crop, but late and early varieties exist. 

2 Honzo komoku keitno, Ch. 18, p. 2. 

8 In a subsequent study on the plants and agriculture of the Indo-Chinese, I 
hope to demonstrate that the Indo-Chinese nations, especially the Chinese and 
Tibetans, possess a common designation for "hemp," and that hemp has been 
cultivated by them in a prehistoric age. There also the history of hemp will be 

294 Sino-Iranica 

(Iranians and Indo-Aryans) possess an identical word for "hemp" (Avestan 
bangha, Sanskrit bhanga), while the European languages have a distinct 
designation, which is presumably a loan-word pointing to Finno-Ugrian 
and Turkish; and, second, that there is a common Old-Turkish word 
for "hemp" of the type kandir, which stands in some relation to the 
Finno-Ugrian appellations. 1 It is most likely that the Scythians brought 
hemp from Asia to Europe. 2 On the other hand, it is well known what 
vital importance flax and linen claimed in the life of the Egyptians 
and the classical peoples. 3 Flax is the typically European, hemp the 
typically Asiatic textile. Surely Linum usitatissimum was known in 
ancient Iran and India. It was and is still wild in the districts included 
between the Persian Gulf, the Caspian Sea, and the Black Sea.* It 
was probably introduced into India from Iran, but neither in India nor 
in Iran wa6 the fibre ever used for garments: the plant was only culti- 
vated as a source of linseed and linseed-oil. 5 Only a relatively modern 
utilization of flax-fibres for weaving is known from a single locality in 
Persia, — Kazirtin, in the province of Fars. This account dates from the 
beginning of the fourteenth century, and the detailed description 
given of the process testifies to its novelty and exceptional character. 6 
This exception confirms the rule. The naturalization of Linum in China, 
of course, is far earlier than the fourteenth century. As regards the 
utilization of Linum, the Chinese fall in line with Iranians and Indo- 
Aryans; and it is from Iranians that they received the plant. The 
case is a clear index of the fact that the Chinese never were in direct 
contact with the Mediterranean culture-area, and that even such culti- 
[ vated plants of this area as reached them were not transmitted from 
there directly, but solely through the medium of Iranians. The case 
is further apt to illustrate how superficial, from the viewpoint of tech- 
nical culture, the influence of the Greeks on the Orient must have 
been since Alexander's campaign, as an industry like flax-weaving 
was not promoted by them, although the material was offered there 
by nature. 

For botanical reasons it is possible that Linum usitatissimum was 
introduced into China from Fergana. There it is still cultivated, and 
only for the exclusive purpose of obtaining oil from the seeds. 7 As has 

1 Z. Qombocz, Bulgarisch-turkische Lehnworter, p. 92. 

2 Cf. for the present, A. de Candolle, Origin of Cultivated Plants, p. 148. 

* Pliny, xix, 1-3; H. Blumner, Technologie, Vol. I, 2d ed., p. 191. 

* A. de Candolle, Origin of Cultivated Plants, p. 130. 

5 See the interesting discussion of Watt, Commercial Products of India, p. 721. 
8 G. Le Strange, Description of the Province of Fars in Persia, p. 55. 
7 S. Korzinski, Vegetation of Turkistan (in Russian), p. 51. 

Sesame and Flax 295 

been pointed out, the plant is indigenous also in northern Persia, and 
must have been cultivated there from ancient times, although we have 
no information on this point from either native documents or Greek 
authors. 1 

Bretschneider 2 says that "flax was unknown to the ancient 
Chinese; it is nowadays cultivated in the mountains of northern China 
(probably also in other parts) and in southern Mongolia, but only for 
the oil of its seeds, not for its fibres; the Chinese call it hu ma ('foreign 
hemp'); the Pen ts'ao does not speak of it; its introduction must be of 
more recent date." This is erroneous. The Pen ts'ao includes this 
species under the ambiguous term hu ma; and, although the date of the 
introduction cannot be ascertained, the event seems to have taken 
place in the first centuries of our era. 

At present, the designation hu ma appears to refer solely to flax. 
A. Henry 5 states under this heading, "This is flax (Linum usitatis- 
simum), which is cultivated in San-si, Mongolia, and the mountainous 
parts of Hu-pei and Se-c'wan. In the last two provinces, from personal 
observation, flax would seem to be entirely cultivated for the seeds, 
which are a common article in Chinese drug-shops, and are used locally 
for their oil, utilized for cooking and lighting purposes." In another 
paper, 4 the same author states that Linum usitatissimum is called at 
Yi-£'an, Se-£'wan, San U ma tfj Be 5 M ("mountain sap-hemp"), and 
that it is cultivated in the mountains of the Patun district, not for the 
fibre, but for the oil which the seed yields. 

Chinese hu ma has passed into Mongol as xuma (khuma) with the 
meaning "sesame," 6 and into Japanese as goma, used only in the sense 
of Sesamum indicum? while Linum usitatissimum is in Japanese ama 
or iUnen-ama} 

Yao Mih-hwi #fc ^ jt?, in his book on Mongolia (Mon-ku &'),* 
mentions hu ma among the products of that country. There are several 
wild-growing species of Linum in northern China and Japan, — ya ma 

1 Joret, Plantes dans l'antiquite\ Vol. II, p. 69. 

2 Bot. Sin., pt. II, p. 204. 

* Chinese Jute, p. 6 (publication of the Chinese Maritime Customs, Shanghai, 

4 Chinese Names of Plants, p. 239 (Journal China Branch Royal As. Soc, 
Vol. XXII, 1887). 

6 The popular writing ;J£, according to the Pen ts'ao kan tnu, is incorrect. 
4 Kovalevski, Dictionnaire mongol, p. 934. 

7 Matsumura, No. 2924. 

8 Ibid., No. 1839. 

9 Ch. 3, p. 41 (Shanghai, 1907). 

296 Sino-Iranica 

jS ffc (Japanese nume-goma or aka-goma), Linum perenne, and Japanese 
matsuba-ninjin or matsuba-nadeUko, Linum possarioides. 1 Forbes and 
Hemsley, 2 moreover, enumerate Linum nutans for Kan-su, and L. 
stelleroides for Ci-li, San-tun, Manchuria, and the Korean Archipelago. 
In northern China, Linum sativum (San-si hu ma \U ]§ #3 M ) is 
cultivated for the oil of its seeds. 3 

1 Matsumura, Nos. 1837, 1838; Stuart, Chinese Materia Medica, p. 242. 

2 Journal Linnean Soc, Vol. XXIII, p. 95. 

8 This species is figured and described in the Ci wu min H t'u k'ao. 


8. The Po wu U, faithful to its tendencies regarding other Iranian 
plants, generously permits General Can K'ien to have also brought back 
from his journey the coriander, hu swi tft ^ (Coriandrum sativum). 1 
Li Si-cen, and likewise K'an-hi's Dictionary, repeat this statement 
without reference to the Po wu li* and of course the credulous com- 
munity of the Changkienides has religiously sworn to this dogma. 3 
Needless to say that nothing of the kind is contained in the General's 
biography or in the Han Annals. 4 The first indubitable mention of the 
plant is not earlier than the beginning of the sixth century a.d.; that 
is, about six centuries after the General's death, and this makes some 
difference to the historian. 5 The first Pen ts'ao giving the name hu-swi 
is the Si liao pen ts'ao, written by Mon Sen in the seventh century, 
followed by the Pen ts'ao H i of C'en Ts'an-k'i in the first half of the 
eighth century. None of these authors makes any observation on 
foreign introduction. In the literature on agriculture, the cultivation 
of the coriander is first described in the Ts'i tnin yao lu of the sixth 
century, where, however, nothing is said about the origin of the plant 
from abroad. 

An interesting reference to the plant occurs in the Buddhist dic- 
tionary Yi ts'ie kin yin i (I.e.), where several variations for writing 

1 This passage is not a modern interpolation, but is of ancient date, as it is cited 
in the Yi ts'ie kin yin i, Ch. 24, p. 2 (regarding this work, see above, p. 258). Whether 
it was contained in the original edition of the Po wu li, remains doubtful. 

1 Under $J ("garlic") K'an-hi cites the dictionary "Tan yun, published by Sun 
Mien in a.d. 750, as saying that the coriander is due to Can K'ien. 

3 Bretschneider, Chinese Recorder, 1871, p. 221, where the term hu-swi is 
wrongly identified with parsley, and Bot. Sin., pt. 1, p. 25; Hirth, T'oung Pao, 
Vol. VI, 1895, p. 439. 

4 The coriander is mentioned in several passages of the Kin kwei yao lio by 
the physician Can Cun-kin of the second century a.d.; but, as stated above (p. 205), 
there is no guaranty that these passages belonged to the original edition of the 
work. "To eat pork together with raw coriander rots away the navel" (Ch. c, 
p. 23 b). "In the fourth and eighth months do not eat coriander, for it injures the 
intellect " (ibid., p. 28). "Coriander eaten for a long time makes man very forgetful; 
a patient must not eat coriander or hwan-hwa ts'ai 3f ^6 ^j| (Latnpsana 
apogonoides) ," ibid., p. 29. 

6 An incidental reference to hu swi is made in the Pen ts'ao kan tnu in 
the description of the plant kuan er (see Bretschneider, Bot. Sin., pt. II, 
No. 438), and ascribed to Lu Ki, who lived in the latter part of the third century 
a.d. In my opinion, this reading is merely due to a misprint, as there is preserved no 
description of the hu-swi by Lu Ki. 


298 Sino-Iranica 

the character swi are given, also the synonymes hian ts'ai # £j| 
("fragrant vegetable") and hian siin # %. 1 In Kian-nan the plant 
was styled hu swi #1 ||j, also hu ki m f§, the pronunciation of the 
latter character being explained by jffi k K i) *gi. The coriander belongs 
to the five vegetables of strong odor (p. 303) forbidden to the geomancers 
and Taoist monks. 2 

I have searched in vain for any notes on the plant that might 
elucidate its history or introduction; but such do not seem to exist, 
not even in the various Pen ts'ao. As regards the Annals, I found only 
a single mention in the Wu Tai Si, 3 where the coriander is enumerated 
among the plants cultivated by the Uigur. In tracing its foreign origin, 
we are thrown back solely on the linguistic evidence. 

The coriander was known in Iran: it is mentioned in the Bundahisn. 4 
Its medical properties are discussed in detail by Abu Mansur in his 
Persian pharmacopoeia. 6 Schlimmer 6 observes, "Se cultive presque 
partout en Perse comme plante potagere; les indigenes le croient 
antiaphrodisiaque et plus sp£cialement an^antissant les ejections." It 
occurs also in Fergana. 7 It was highly appreciated by the Arabs in their 
pharmacopoeia, as shown by the long extract devoted to it by Ibn 
al-Baitar. 8 In India it is cultivated during the cold season. The San- 
skrit names which have been given on p. 284, mean simply "grain," 
and are merely attributes, 9 not proper designations of the plant, for 
which in fact there is no genuine Sanskrit word. As will be seen below, 
Sanskrit kustumburu is of Iranian origin; and there is no doubt in my 
mind that the plant came to India from Iran, in the same manner as 
it appears to have spread from Iran to China. 

j^ |£ or |g hu-swi, *ko(go)-swi (su), appears to be the transcription 
of an Iranian form *koswi, koswi, goswi. Cf. Middle Persian golniz; 

1 Two dictionaries, the Tse yuan ^r %L and Yiin Ho fH B§, are quoted in this 
text, but their date is not known to me. As stated in the Pen ts ' ao si i and Si wu ki yuan 
(Ch. 10, p. 30; above, p. 279), the change from hu swi to hian swi was dictated by a taboo 
imposed by Si Lo ^ ^ (a.d. 273-333), who was himself a Hu (cf. below, 
p. 300) ; but we have no contemporaneous account to this effect, and the attempt 
at explanation is surely retrospective. 

* Pen ts'ao kan mu, Ch. 26, p. 6 b; and Stuart, Chinese Materia Medica, p. 28. 

« Ch. 74, p. 4. 

4 Above, p. 192. 

6 Achundow, Abu Mansur, p. 112. 

6 Terminologie, p. 156. 

7 S. Korzinski, Vegetation of Turkistan (in Russian), p. 51. 

8 L. Leclerc, Traits des simples, Vol. Ill, pp. 170-174. 

9 Such are also the synonymes suksmapatra, tiksnapatra, tiksnapkala ("with 
leaves or fruits of sharp taste"). 

The Cortander 299 

New Persian kis'niz,kus'niz, and gilnlz, also Siiniz; 1 Kurd ksnis or ktfniS; 
Turkish kiSnii; Russian kiSnits; Aramaic kusbarta and kusbar (Hebrew 
gad, Punic yolS, are unconnected), Arabic kozbera or kosher et; Sanskrit 
kustumburu and kustumbari; Middle and Modern Greek Kowtapas 1 
and ki<jvvt)t$i. 

According to the Hut k'ian ci, the coriander is called in Turkistan 
(that is, in Turkl) yun-ma-su f& M ^. 

It is commonly said that the coriander is indigenous to the Mediter- 
ranean and Caucasian regions (others say southern Europe, the Levant, 
etc.), but it is shown by the preceding notes that Iran should be included 
in this definition. I do not mean to say, however, that Iran is the ex- 
clusive and original home of the plant. Its antiquity in Egypt and in 
Palestine cannot be called into doubt. It has been traced in tombs of 
the twenty-second dynasty (960-800 B.C.), 3 and Pliny 4 states that the 
Egyptian coriander is the best. In Iran the cultivation seems to have 
been developed to a high degree; and the Iranian product was propa- 
gated in all directions, — in China, India, anterior Asia, and Russia. 

The Tibetan name for the coriander, u-su, may be connected with 
or derived from Chinese hu-sui. L. A. Waddell 6 saw the plant culti- 
vated in a valley near Lhasa. It is also cultivated in Siam. 6 

Coriander was well known in Britain prior to the Norman Con- 
quest, and was often employed in ancient Welsh and English medicine 
and cookery. 7 Its Anglo-Saxon name is cellendre, coliandre, going back 
to Greek koridndron, koriannon. / 

1 Another Persian word is bughunj. According to Steingass (Persian Diction- 
ary), talki or tdlgi denotes a "wild coriander." 

' The second element of the Arabic, Sanskrit, and Greek words seems to bear 
some relation to Coptic berSiu, herein (V. Loret, Flore pharaonique, p. 72). In 
Greece, coriander is still cultivated, but only sparsely, near Theben, Corinth, and 
Cyparissia (Th. v. Heldreich, Nutzpflanzen Griechenlands, p. 41). 

* V. Loret, op. cit., p. 72; F. Woenig, Pflanzen im alten Aegypten, p. 225. 
4 xx, 20, §82. 

* Lhasa, p. 316. 

* Pallegoix, Description du royaume thai, Vol. I, p. 126. 
7 FlOckiger and Hanbury, Pharmacographia, p. 329. 


9. Another dogma of the Changkienomaniacs is that the renowned 
General should have also blessed his countrymen with the introduction 
of the cucumber (Cucumis sativus), styled hu kwa $j jU. ("Iranian 
melon") or hwan kwa lit JR ("yellow melon"). 1 The sole document 
on which this opinion is based is presented by the recent work of Li 
Si-Sen, 2 who hazards this bold statement without reference to any older 
authority. Indeed, such an earlier source does not exist: this bit of 
history is concocted ad hoc, and merely suggested by the name hu kwa. 
Any plants formed with the attribute hu were ultimately palmed off on 
the old General as the easiest way out of a difficult problem, and as a 
comfortable means of saving further thought. 

Li Si-cen falls back upon two texts only of the T'ang period, — the 
Pen ts'ao H i, which states that the people of the north, in order to avoid 
the name of Si Lo 15 Wi (a.d. 273-333), wno was oi Hu descent, tabooed 
the term hu kwa, and replaced it by hwan kwa; 3 and the Si i lu #Hft§ifc 
by Tu Pao $1 H, who refers this taboo to the year 608 (fourth year 
of the period Ta-ye of the Sui dynasty). 4 If this information be correct, 
we gain a chronological clew as to the terminus a quo: the cucumber 
appears to have been in China prior to the sixth century a.d. Its culti- 
vation is alluded to in the Ts'i min yao iw from the beginning of the 
sixth century, provided this is not an interpolation of later times. 5 

According to Engler, 8 the home of the cucumber would most prob- 

1 Bretschneider, Chinese Recorder, 1871, p. 21 (accordingly adopted by 
de Candolle, Origin of Cultivated Plants, p. 266); Stuart, Chinese Materia 
Medica, p. 135. In Japanese, the cucumber is ki-uri. 

2 Pen ts'ao kan mu, Ch. 28, p. 5 b. 

8 A number of other plant-names was hit by this taboo (cf . above, p. 298) : thus 
the plant lo-lo f^ ^ (Ocimum basilicum), which bears the same character as SiLo's 
personal name, as already indicated in the Ts'i min yao Su (see also Si wu ki yuan, 
Ch. 10, p. 30 b; Ci wu min Si t'u k'ao, Ch. 5, p. 34; and Pen ts'ao kan mu, Ch. 26, 
p. 22 b). He is said to have also changed the name of the myrobalan ho-li-lo (below, 
p. 378) into ho-tse fSf J*. There is room for doubt, however, whether any of these 
plants existed in the China of his time; the taboo explanations may be makeshifts 
of later periods. 

4 This is the Ta ye Si i lu (Records relative to the Ta-ye period, 605-618)1 
mentioned by Bretschneider (Bot. Sin., pt. 1, p. 195). The Pen ts'ao kan mu 
(Ch. 22, p. 1) quotes the same work again on the taboo of the term hu ma (p. 288), 
which in 608 was changed into kiao ma ^ j§jc. 

6 Cf. Ci wu min Si t'u k'ao, Ch. 5, p. 43. 

6 In Hehn, Kulturpflanzen, p. 323. 


The Cucumber 301 

; ably be in India; and Watt 1 observes, "There seems to be no doubt 
■' that one at least of the original homes of the cucumber was in North 
, India, and its cultivation can be traced to the most ancient classic times 
of Asia." De Candolle 2 traces the home of the plant to northwestern 
India. I am not yet convinced of the correctness of this theory, as the 
historical evidence in favor of India, as usual in such cases, is weak; 3 
and the cultivation of the cucumber in Egypt and among the Semites 
is doubtless of ancient date. 4 At any rate, this Cucurbitacea belongs to 
the Egypto- West-Asiatic culture-sphere, and is not indigenous to 
China. There is, however, no trace of evidence for the gratuitous 
speculation that its introduction is due to General Can K'ien. The 
theory that it was transmitted from Iranian territory is probable, but 
there is thus far no historical document to support it. The only trace 
of evidence thereof appears from the attribute Hu. 

Abu Mansur mentions the cucumber under the name qittd, adding 
the Arabic-Persian xiyar and kawanda in the language of Khorasan. 5 
The word xiyar has been adopted into Osmanli and into Hindustani in 
the form xird. Persian xdwuS or xawaS denotes a cucumber kept for 
seed; it means literally "ox-eye" (gav-aS; Avestan a$i, Middle Persian 
ai, Sanskrit ak$i, "eye"), corresponding to Sanskrit gavdk$i ("a kind 
of cucumber"). A Pahlavi word for "cucumber" is vdtrah, which 
developed into New Persian bddran, bdlah, or varan (Afghan bddran). 6 

1 Commercial Products of India, p. 439. In Sanskrit the cucumber is trapusa. 

* Op. til., p. 265. 

3 Such a positive assertion as that of de Candolle, that the cucumber was 
cultivated in India for at least three thousand years, cannot be accepted by any 
serious historian. 

* V. Loret, Flore pharaonique, p. 75; C. Joret, Plantes dans l'antiquitd, 
Vol. I, p. 61. 

6 Achundow, AbujMansur, p. 106. 

6 This series is said to mean also "citron." The proper Persian word for the 
latter fruit is turuttj (Afghan turanj, Balu& trunj). The origin of this word, as far 
as I know, has not yet been correctly explained, not even by HObschmann (Armen . 
Gram., p. 266). Vullers (Lexicon persico-latinum, Vol. I, p. 439) tentatively 
suggests derivation from Sanskrit suranga, which is surely impossible. The real 
source is presented by Sanskrit matulunga ("citron," Citrus medico). 


10. Although a number of alliaceous plants are indigenous to China, 1 
there is one species, the chive (Allium scorodoprasum; French rocambole), 
to which, as already indicated by its name hu swan #1 m or hu j$J 
("garlic of the Hu, Iranian garlic"), a foreign origin is ascribed by the 
Chinese. Again, the worn-out tradition that also this introduction 
is due to Can K'ien, is of late origin, and is first met with in the 
spurious work Po wu U, and then in the dictionary T'an yiin of the middle 
of the eighth century. 2 Even Li Si-oen 3 says no more than that "people 
of the Han dynasty obtained the hu swan from Central Asia." It seems 
difficult, however, to eradicate a long-established prejudice or an error 
even from the minds of scholars. In 191 5 I endeavored to rectify it, 
especially with reference to the wrong opinion expressed by Hirth in 
1895, that garlic in general must have been introduced into China 
for the first time by Can K'ien. Nevertheless the same misconception 
is repeated by him in 191 7, 4 while a glance at the Botanicon Sinicum 6 
would have convinced him that at least four species of Allium are of 
a prehistoric antiquity in China. The first mention of this Central- 
Asiatic or Iranian species of Allium is made by T'ao Huh-kifi 
(a.d. 45 1-536) , provided the statement attributed to him in the Ceh lei pen 
ts'ao and Pen ts'ao kah mu really emanates from him. 6 When the new 
Allium was introduced, the necessity was felt of distinguishing it from the 
old, indigenous Allium sativum, that was designated by the plain root- 
word swan. The former, accordingly, was characterized as ta swan 
^C ffii ("large Allium"); the latter, as siao /h swan ("small Allium"). 
This distinction is said to have first been recorded by T'ao Huh-kin. 
Also the Ku kin lu is credited with the mention of hu swan; this, how- 
ever, is not the older Ku kin cu by Ts'ui Pao of the fourth century, but, 
as expressly stated in the Pen ts'ao, the later re-edition by Fu Hou 

1 Cf. Voung Pao, 1915, pp. 96-99. 

2 Bretschneider, Bot. Sin., pt. Ill, No. 244. 

* Pen ts'ao kan mu, Ch. 26, p. 6 b. 

* Journal Am. Or. Soc, Vol. XXXVII, p. 92. 

6 Pt. II, Nos. 1-4, 63, 357-360, and III, Nos. 240-243. 

6 The Kin kwei yao Ho (Ch. c, p. 24 b) of the second century a.d. mentions hu 
swan, but this in all probability is a later interpolation (above, p. 205). 


Chive, Onion, and Shallot 303 

{£ $£ of the tenth century. However, this text is now inserted in the 
older Ku kin lu, x which teems with interpolations. 

Ta swan is mentioned also as the first among the five vegetables of 
strong odor tabooed for the Buddhist clergy, the so-called wu hun 
3l j|£. 2 This series occurs in the Brahmajala-sutra, translated in 
a.d. 406 by Kumarajlva. 8 If the term ta swan was contained in the 
original edition of this work, we should have good evidence for carry- 
ing the date of the chive into the Eastern Tsin dynasty (a.d. 317-419). 

11. There is another cultivated species of Allium (probably A. 
fistulosum) derived from the West. This is first mentioned by Sun Se- 
miao M B j!!, 4 in his Ts'ien kin H U ^ %. "% 7n (written in the begin- 
ning of the seventh century), under the name hu ts'un i§ ^, because 
the root of this plant resembles the hu swan iPf |jf. It was usually styled 
swan-ts'un m i§? or hu #1 ts'un (the latter designation in the K'ai poo 
pen ts'ao of the Sung). In the Yin San Zen yao (p. 236), written in 1331 
under the Yuan, it is called hui-hui ts'un 01 ("Mohammedan 
onion"). 6 This does not mean, however, that it was only introduced 
by Mohammedans; but this is simply one of the many favorite alter- 
ations of ancient names, as they were in vogue during the Mongol 
epoch. This Allium was cultivated in Se-£'wan under the T'ang, as 
stated by Mon Sen j£ $9g in his Si liao pen ts'ao, written in the second 
half of the seventh century. Particulars in regard to the introduction 
are not on record. 

12. There is a third species of Allium, which reached China under 
the T'ang, and which, on excellent evidence, may be attributed to 
Persia. In a.d. 647 the Emperor T'ai Tsuri solicited from all his tribu- 
tary nations their choicest vegetable products, 6 and their response to 
the imperial call secured a number of vegetables hitherto unknown in 
China. One of these is described as follows: "Hun-t'i onion $? $| ^ 
resembles in appearance the onion (ts'un, Allium fistulosum), but is 
whiter and more bitter. On account of its smell, it serves as a remedy. 

1 Ch. c, p. 3 b. 

2 This subject is treated in the Pen ts'ao kan mu (Ch. 26, p. 6 b) under the 
article swan, and summed up by Stuart (Chinese Materia Medica, p. 28). See, 
further, De Groot, Le Code du Mahayana en Chine, p. 42, where the five plant- 
names are unfortunately translated wrongly (hin-k'u, "asafcetida" [see p. 361], is 
given an alleged literal translation as "le lys d'eau montant"!), and Chavannes 
and Pelliot, Traits manicheen, pp. 233-235. 

• Bunyiu Nanjio, Catalogue of the Buddhist Tripitaka, No. 1087. 

4 Cf. below, p. 306. 

6 Pen ts'ao kan mu, Ch. 26, p. 5. 

8 We shall come back to this important event in dealing with the history of the 

304 Sino-Iranica 

In its appearance it is like lan-lin-tun M §1 4K 1 but greener. When 
dried and powdered, it tastes like cinnamon and pepper. The root is 
capable of relieving colds." 2 The Fun H wen kien ki 3 adds that hun-fi 
came from the Western Countries (Si yii). 

Hun-Vi is a transcription answering to ancient *gwun-de, and 
corresponds to Middle Persian gandena, New Persian gandand, Hindi 
gandand, Bengali gundina (Sanskrit mleccha-kanda, "bulb of the bar- 
barians"), possibly the shallot (Allium ascalonicum; French ichalotte, 
ciboule) or A. porrum, which occurs in western Asia and Persia, but not 
in China. 4 

Among the vegetables of India, Huan Tsan 5 mentions W K5 hun-Vo 
(*hun-da) ts'ai. Julien left this term untranslated; Beal did not know, 
either, what to make of it, and added in parentheses kandu with an 
interrogation-mark. Waiters 6 explained it as "kunda (properly the 
olibanum-tree)." This is absurd, as the question is of a vegetable culti- 
vated for food, while the olibanum is a wild tree offering no food. More- 
over, hun cannot answer to kun; and the Sanskrit word is not kunda, 
but kundu or kunduru. The mode of writing, hun, possibly is intended 
to allude to a species of Allium. Huan Tsan certainly transcribed a 
Sanskrit word, but a Sanskrit plant-name of the form kunda or gunda 
is not known. Perhaps his prototype is related to the Iranian word 
previously discussed. 

1 The parallel text in the Ts'e fu yuan kwei (Ch. 970, p. 12) writes only lin-tun. 
This plant is unidentified. 

2 Tan hui yao, Ch. 100, p. 3 b; and Ch. 200, p. 14 b. 

3 Ch. 7, p. 1 b (above, p. 232). 

4 A. de Candolle, Origin of Cultivated Plants, pp. 68-71; Leclerc, Traite" 
des simples, Vol. Ill, pp. 69-71; Achundow, Abu Mansur, pp. 113, 258. Other 
Persian names are tara and kawar. They correspond to Greek xpdaov, Turkish 
prasa, Arabic karat. The question as to whether the species ascalonicum or porrum 
should be understood by the Persian term gandand, I have to leave in suspense and 
to refer to the decision of competent botanists. Schlimmer (Terminologie, p. 21) 
identifies Persian gandand with Allium porrum; while, according to him, A. ascalon- 
icum should be musir in Persian. Vullers (Lexicon persico-latinum, Vol. II, p. 1036) 
translates the word by "porrum." On the other hand, Stuart (Chinese Materia 
Medica, p. 25), following P. P. Smith, has labelled Chinese hiai $£, an Allium 
anciently indigenous to China, as A. ascalonicum. If this be correct, the Chinese 
would certainly have recognized the identity of the foreign hun-t'i with hiai, provided 
both should represent the same species, ascalonicum. Maybe also the two were 
identical species, but differentiated by cultivation. 

6 Ta Tan si yii ki, Ch. 2, p. 8 b. 

6 On Yuan Chwang's Travels, Vol. I, p. 178. 


13. Among the many species of pulse cultivated by the Chinese, 
there are at least two to which a foreign origin must be assigned. Both 
are comprised under the generic term hu tou ffli 5L ("bean of the Hu," 
or "Iranian bean"), but each has also its specific nomenclature. It 
is generally known that, on account of the bewildering number of species 
and variations and the great antiquity of their cultivation, the history 
of beans is fraught with graver difficulties than that of any other group 
of plants. 

The common or garden pea (Pisum sativum) is usually styled wan 
tou *&& .3. (Japanese Hro-endo), more rarely ts'in siao tou W 4* 3. 
("green small pulse"), tsHn pan tou W ?£ .2. ("green streaked pulse"), 
and ma lei M%. A term ^ SL pi tou, *pit (pir) tou, is regarded as 
characteristic of the T'ang period; while such names as hu tou, 2un $u 
r^L^L ("pulse of the Zuh"), 1 and hui-hu toHWl ("pulse of the 
Uigur;" in the Yin San fen yao of the Mongol period changed also into 
hui-hui tou 5., "Mohammedan pulse") are apt to bespeak the 
foreign origin of the plant. 2 Any document alluding to the event of the 
introduction, however, does not appear to exist in Chinese records. 
The term hu tou occurs in the present editions of the Ku kin ?u, 3 hu-la 
Oft & being given as its synonyme, and described as "resembling the 
li touV£.3~, but larger, the fruit of the size of a child's fist and eatable." 
The term li tou is doubtfully identified with Mucuna capitata;* but the 
species of the Ku kin lu defies exact identification; and, as is well known, 
this book, in its present form, is very far from being able to claim abso- 
lute credence or authenticity. Also the Kwan li, written prior to 
a.d. 5 2 7, contains the term hu tou; 6 but this name, unfortunately, is ambig- 
uous. Li Si-cen acquiesces in the general statement that the pea has 
come from the Hu and Zuh or from the Western Hu (Iranians) ; he cites, 
however, a few texts, which, if they be authentic, would permit us to 

1 This term is ambiguous, for originally it applies to the soy-bean (Glycine 
hispida), which is indigenous to China. 

1 Cf. Pen ts'ao kan mu, Ch. 24, p. 7; and Kwan k'un fan p'u, Ch. 4, p. II. The 
list of the names for the pea given by Bretschneider (Chinese Recorder, 1871, 
p. 223) is rather incomplete. 

8 Ch. b, p. 1 b. 

4 Stuart, Chinese Materia Medica, p. 269. The word /* is also written ^. 

s Tai p'in yu Ian, Ch. 841, p. 6 b. 


306 Sino-Iranica 

fix approximately the date as to when the pea became known to the 
Chinese. Thus he quotes the Ts'ien kin fan X 4£ ^ of the Taoist 
adept Sun Se-miao $£ JS jII, 1 of the beginning of the seventh century, as 
mentioning the term hu tou with the synonymes tsHh siao tou and ma-lei. 
The Ye Zuh ki 2 of the fourth century a.d. is credited with the statement 
that, when Si Hu tabooed the word hu t$l, the term hu tou was altered 
into kwo tou 10 2. ("bean of the country," "national bean"). Accord- 
ing to Li Si-cen, these passages allude to the pea, for anciently the 
term hu tou was in general use instead of wan tou. He further refers to 
the T'an H M it as saying that the pi tou comes from the Westerta 
2un and the land of the Uigur, and to the dictionary Kwan ya by Can 
Yi (third century a.d.) as containing the terms pi tou, wan tou, and liu 
tou i§ it. It would be difficult to vouchsafe for the fact that these 
were really embodied in the editio princeps of that work; yet it would 
not be impossible, after all, that, like the walnut and the pomegranate, 
so also the pea made its appearance on Chinese soil during the fourth 
century a.d. There can be no doubt of the fact that it was cultivated in 
China under the T'ang, and even under the Sui (a.d. 590-617). In the 
account of Liu-kiu (Formosa) it is stated that the soil of the island is 
advantageous for the cultivation of hu tou. 3 Wu K'i-tsun 4 contradicts 
Li Si-cen's opinion, stating that the terms hu tou and wan tou apply to 
different species. 

None of the Chinese names can be regarded as the transcription of 
an Iranian word. Pulse played a predominant part in the nutrition of 
Iranian peoples. The country Si (Tashkend) had all sorts of pulse. 5 
Abu Mansur discusses the pea under the Persian name xulldr and the 
Arabic julban. 6 Other Persian words for the pea are nujud and gergeru 
or xereghan. 1 

A wild plant indigenous to China is likewise styled hu tou. It is 
first disclosed by C'en Ts'an-k'i of the T*ang period, in his Pen ts'ao H i, 
as growing wild everywhere in rice-fields, its sprouts resembling the 
bean. In the Ci wu mih H Vu k'ao 8 we meet illustrations of two wild 

1 Regarding this author, see Wylie, Notes on Chinese Literature, pp. 97, 99; 
Bretschneider, Bot. Sin., pt. 1, p. 43; L. Wieger, Taoisme, le canon, pp. 142, 143, 
182; Pelliot, Bull, de I'Ecolefrangaise, Vol. IX, pp. 435-438. 

* See above, p. 280. 
*SuiSu, Ch. 81, p. 5 b. 

* Ci wu tnin Si Vu k'ao, Ch. 2, p. 150. 
8 T'ai p'in hwan yii ki, Ch. 186, p. 7 b. 

6 Achundow, Abu Mansur, pp. 41, 223. 

T The latter is given by Schlimmer (Terminologie, p. 464). 

8 Ch. 2, pp. 11, 15. 

Garden Pea and Broad Bean 307 

plants. One is termed hui-hui tou ("Mohammedan bean"), first men- 
tioned in the Kiu hwan pen ts'ao of the fourteenth century, called also 
na-ho tou ffi & JUL, the bean being roasted and eaten. The other, 
named hu tou, is identified with the wild hu tou of C'en Ts'ah-k'i; and 
Wu K'i-tsun, author of the Ci wu min H t'u k'ao, adds the remark, 
"What is now called hu tou grows wild, and is not the hu tou [that is, 
the pea] of ancient times." 

14. On the other hand, the term hu tou t® S. refers also to Faba 
sativa (F. vulgaris, the vetch or common bean), according to Bret- 
schneider, 1 "one of the cultivated plants introduced from western 
Asia into China, in the second century B.C., by the famous general 
Chang K'ien." This is an anachronism and a wild statement, which he 
has not even supported by any Chinese text. 2 The history of the species 
in China is lost, or was never recorded. The supposition that it was 
introduced from Iran is probable. It is mentioned under the name 
pag (gavirs) in the Bundahisn as the chief of small-seeded grains.* 
Abu Mansur has it under the Persian name baqild or bdqla.* Its culti- 
vation in Egypt is of ancient date. 6 

15. Ts'an tou MIL ("silkworm bean," so called because in its 
shape it resembles an old silkworm), Japanese sorantame, the kidney- 
bean or horse-bean (Vicia faba), is also erroneously counted by Bret- 
schneider 8 among the Can-K'ien plants, without any evidence being 
produced. It is likewise called hu tou tft &,, but no historical documents 
touching on the introduction of this species are on record. It is not 
mentioned in T'ang or Sung literature, and seems to have been intro- 
duced not earlier than the Yuan period (1260-1367). It is spoken of 
in the Nun Su §k # ("Book on Agriculture") of Wan Ceri 3: H of 
that period, and in the Kiu hwan pen ts'ao $C %L ^ P of the early 

1 Bot. Sin., pt. II, No. 29. 

• The only text to this effect that I know of is the Pen ts'ao kin, quoted in the 
Pat p'in yii Ian (Ch. 841, p. 6 b), which ascribes to Can K'ien the introduction of 
sesame and hu tou; but which species is meant (Pisum sativum, Faba sativa, or 
Vicia faba) cannot be guessed. The work in question certainly is not the Pen ts'ao 
kin of Sen-nun, but it must have existed prior to a.d. 983, the date of the publication 
of the Pet p'in yii Ian. 

• West, Pahlavi Texts, Vol. I, p. 90. 
4 Achundow, Abu Mansur, p. 20. 

' V. Loret, Flore pharaonique, p. 94. 

• Chinese Recorder, 1 871, p. 221 (thus again reiterated by de Candolle, Origin 
of Cultivated Plants, p. 318). The Kwan k'iin fan p'u (Ch. 4, p. 12 b) refers the 
above text from the T'ai p'in yii Ian to this species, but also to the pea. This con- 
fusion is hopeless. 

308 Sino-Iranica 

Ming, 1 which states that "now it occurs everywhere." Li Si-cen says 
that it is cultivated in southern China and to a larger extent in Se- 
6'wan. Wan Si-mou HE tfr $8i, who died in 1591, in his Hio pu tsa $u 
^ IS ft $M, a work on horticulture in one chapter, 2 mentions an espe- 
cially large and excellent variety of this bean from Yun-nan. This is 
also referred to in the old edition of the Gazetteer of Yun-nan Province 
(Kiu Yun-nan Vun U) and in the Gazetteer of the Prefecture of Mun- 
hwa in Yun-nan, where the synonyme nan tou l$J s£ ("southern bean") 
is added, as the flower turns its face toward the south. The New-Persian 
name of the plant is bagela. 3 

1 Ci wu min H t'u k'ao, Ch. 2, p. 142. Bretschneider (Bot. Sin., pt. 1, p. 52) 
has recognized Vicia faba among the illustrations of this work. 

1 Cf. the Imperial Catalogue, Ch. 116, p. 37 b. 

3 Schlimmer, Terminologie, p. 562. Arabic b&qil&. Finally, the Fan yi min yi tsi 
(section 27) offers a Sanskrit term %fl $jl wu-kia, *mwut-g'a, translated by hu tou 
and explained as "a green bean." The corresponding Sanskrit word is tnudga 
(Phaseolus tnungo), which the Tibetans have rendered as mon sran rdeu, the term 
Mon alluding to the origin from northern India or Himalayan regions (Mint. Soc. 
finno-ougrienne, Vol. XI, p. 96). The Persians have borrowed the Indian word in the 
form mung, which is based on the Indian vernacular munga or mungu (as in Singha- 
lese; Pali mugga). Phaseolus tnungo is peculiar to India, and is mentioned in Vedic 
literature (Macdonell and Keith, Vedic Index, Vol. II, p. 166). 


1 6. Saffron is prepared from the deep orange-colored stigmas, 
with a portion of the style, of the flowers of Crocus sativus (family 
Irideae). The dried stigmas are nearly 3 cm long, dark red, and aro- 
matic, about twenty thousand of them making a pound, or a grain 
containing the stigmas and styles of nine flowers. It is a small plant 
with a fleshy bulb-like conn and grassy leaves with a beautiful purple 
flower blossoming in the autumn. As a dye, condiment, perfume, and 
medicine, saffron has always been highly prized, and has played an 
important part in the history of commerce. It has been cultivated in 
western Asia from remote ages, so much so that it is unknown in a 
wild state. It was always an expensive article, restricted mostly to the 
use of kings and the upper classes, and therefore subject to adulteration 
and substitutes. 1 In India it is adulterated with safflower (Carthamus 
tinctorius), which yields a coloring-agent of the same deep-orange color, 
and in Oriental records these products are frequently confused. Still 
greater confusion prevails between Crocus and Curcuma (a genus of 
Zingiberaceae) , plants with perennial root-stocks, the dried tubers of 
which yield the turmeric of commerce, largely used in the composition 
of curry-powder and as a yellow dye. It appears also that the flowers 
of Memecylon tinctorium were substituted for saffron as early as the 
seventh century. The matter as a subject of historical research is there- 
fore somewhat complex. 

Orientalists have added to the confusion of Orientals, chiefly being 
led astray by the application of our botanical term Curcuma, which is 
derived from an Oriental word originally relating to Crocus, but also 
confounded by the Arabs with our Curcuma. It cannot be too strongly 
emphasized that Sanskrit kunkuma strictly denotes Crocus sativus, 
but never our Curcuma or turmeric (which is Sanskrit haridra), 2 and 

1 Pliny already knew that there is nothing so much adulterated as saffron 
(adulteratur nihil aeque. — xxi, 17, §31). E. Wiedemann (Sitzber. Phys.-tned. 
Sot. ErL, 1914, pp. 182, 197) has dealt with the adulteration of saffron from Arabic 
sources. According to Watt (Commercial Products of India, p. 430), it is too 
expensive to be extensively employed in India, but is in request at princely marriages, 
and for the caste markings of the wealthy. 

2 This is not superfluous to add, in view of the wrong definition of kunkuma 
given by Eitel (Handbook of Chinese Buddhism, p. 80). Sanskrit kdvera ("saffron") 
and kaverl ("turmeric") do not present a confusion of names, as the two words 
are derived from the name of the trading-place Kavera, Chaveris of Ptolemy and 
Caber of Cosmas (see MacCrindle, Christian Topography of Cosmas, p. 367). 


310 Sino-Iranica 

that our genus Curcuma has nothing whatever to do with Crocus or 

As regards Chinese knowledge of saffron, we must distinguish two 
long periods, — first, from the third century to the T'ang dynasty 
inclusive, in which the Chinese received some information about the 
plant and its product, and occasionally tribute-gifts of it; and, second, 
the Mongol period (1260-1367), when saffron as a product was actually 
imported into China by Mohammedan peoples and commonly used. 
This second period is here considered first. 

Of no foreign product are the notions of the Chinese vaguer than 
of saffron. This is chiefly accounted for by the fact that Crocus sativus 
was hardly ever transplanted into their country, 1 and that, although 
the early Buddhist travellers to India caught a glimpse of the plant 
in Kashmir, their knowledge of it always remained rather imperfect. 
First of all, they confounded saffron with safflower (Carthamus tinctori- 
us), as the products of both plants were colloquially styled "red 
flower" {huh hwa %Ljfc). Li Si-cen 2 annotates, "The foreign (fan #) 
or Tibetan red flower [saffron] comes from Tibet (Si-fan) , the places of 
the Mohammedans, and from Arabia (T'ien-fan % 3f). It is the 
huh-lan [Carthamus] of those localities. At the time of the Yuan 
(1 260-1367) it was used as an ingredient in food-stuffs. According to 
the Po wu ci of Can Hwa, Can K'ien obtained the seeds of the huh-lan 
[Carthamus] in the Western Countries (Si yu) , which is the same species 
as that in question [saffron], although, of course, there is some difference 
caused by the different climatic conditions . " It is hence erroneous to state, 
as asserted by F. P. Smith, 3 that "the story of Can K'ien is repeated for 
the saffron as well as for the safflower;" and it is due to the utmost con- 
fusion that Stuart 4 writes, "According to the Pen-ts'ao, Crocus was 
brought from Arabia by Can K'ien at the same time that he brought the 
safflower and other Western plants and drugs." Can K'ien in Arabia! 
The Po wu ci speaks merely of safflower (Carthamus), not of saffron 
(Crocus), — two absolutely distinct plants, which even belong to different 
families; and there is no Chinese text whatever that would link the 
saffron with Can K'ien. In fact, the Chinese have nothing to say re- 

1 It is curious that the Armenian historian Moses of Khorene, who wrote about 
the middle of the fifth century, attributes to China musk, saffron, and cotton (Yule, 
Cathay, Vol. I, p. 93). Cotton was then not manufactured in China; likewise is 
saffron cultivation out of the question for the China of that period. 

2 Pen ts'ao kan mu, Ch. 15, p. 14 b. 

8 Contributions towards the Materia Medica of China, p. 189. 
4 Chinese Materia Medica, p. 131. 

Saffron and Turmeric 311 

garding the introduction or cultivation of saffron. 1 The confusion of 
Li Si-cen is simply due to an association of the two plants known as 
"red flower." Safflower is thus designated in the TsH min yao $u, 
further by Li Cun ^ *r of the T'ang and in the Sun H, where the yen-li 
red flower is stated to have been sent as tribute by the prefecture of 
Hih-yuan % 7C in Sen-si. 2 

The fact that Li Si-cen in the above passage was thinking of 
saffron becomes evident from two foreign words added to his nomen- 
clature of the product: namely, '& i£ Ml ki-ju-lan and fift £fe IP sa-fa- 
tsi. The first character in the former transcription is a misprint for "& 
tsa (*tsap, dzap); the last character in the latter form must be emen- 
dated into &P Ian. 3 Tsa-fu-lan and sa-fa-lan (Japanese sqfuran, Siamese 
faran), as was recognized long ago, represent transcriptions of 
Arabic za'feran or za'fardn, which, on its part, has resulted in our "saf- 

1 Bretschneider (Chinese Recorder, 1871, p. 222) asserts that saffron is not 
cultivated in Peking, but that it is known that it is extensively cultivated in other 
parts of China. I know nothing about this, and have never seen or heard of any 
saffron cultivation in China, nor is any Chinese account to that effect known to me. 
Crocus sativus is not listed in the great work of F. B. Forbes and W. B. Hemsley 
(An Enumeration of All the Plants known from China Proper, comprising Vols. 
23, 26, and 36 of the Journal of the Linnean Society), the most comprehensive syste- 
matic botany of China. Engler (in Hehn, Kulturpflanzen, p. 270) says that Crocus 
is cultivated in China. Watt (Dictionary, Vol. II, p. 593) speaks of Chinese saffron 
imported into India. It is of especial interest that Marco Polo did not find saffron 
in China, but he reports that in the province of Fu-kien they have "a kind of fruit, 
resembling saffron, and which serves the purpose of saffron just as well" (Yule, 
Marco Polo, Vol. II, p. 225). It may be, as suggested by Yule after Fluckiger, that 
this is Gardenia florida, the fruits of which are indeed used in China for dyeing-pur- 
poses, producing a beautiful yellow color. On the other hand, the Pen ts'ao kan mu 
H i (Ch. 4, p. 14 b) contains the description of a "native saffron" (t'u hun hwa j^ 
j£C ^£, in opposition to the "Tibetan red flower" or genuine saffron) after the Con- 
tinued Gazetteer of Fu-kien jjSji %£ $f ^, as follows: "As regards the native 
Saffron, the largest specimens are seven or eight feet high. The leaves are like those 
Df the p'i-P'a flfc |C (Eriobotrya japonica), but smaller and without hair. In the 
autumn it produces a white flower like a grain of maize (su-mi 5!l ^» Zea mays). 
It grows in Fu-cou and Nan-hen-cou f|] ,§> '}\\ [now Yan-kian ££ in Kwah-tuh] 
in the mountain wilderness. That of Fu-cou makes a fine creeper, resembling the 
fu-yun (Hibiscus ntutabilis), green above and white below, the root being like that of 
the ko J§ (Pachyrhizus thunbergianus) . It is employed in the pharmacopoeia, being 
finely chopped for this purpose and soaked overnight in water in which rice has been 
scoured; then it is soaked for another night in pure water and pounded: thus it is 
ready for prescriptions." This species has not been identified, but may well be 
Marco Polo's pseudo-saffron of Fu-kien. 

2 T'u $u tsi ten, XX, Ch. 158. 

3 Cf. Watters, Essays on the Chinese Language, p. 348. This transcription, 
however, does not prove, as intimated by Watters, that "this product was first 
imported into China from Persia direct or at least obtained immediately from 
Persian traders." The word zafardn is an Arabic loan-word in Persian, and may 
have been brought to China by Arabic traders as well. 

312 Sino-Iranica 

fron." 1 It is borne out by the very form of these transcriptions that 
they cannot be older than the Mongol period when the final consonants 
had disappeared. Under the T'ang we should have *dzap-fu-lam and 
*sat-fap-lan. This conclusion agrees with Li Si-Sen's testimony that 
saffron was mixed with food at the time of the Yuan, — an Indo-Persian 
custom. Indeed, it seems as if not until then was it imported and used 
in China; at least, we have no earlier document to this effect. 

Saffron is not cultivated in Tibet. There is no Crocus tibetan us, as 
tentatively introduced by Perrot and Hurrier 2 on the basis of the 
Chinese term "Tibetan red flower." This only means that saffron is 
exported from Tibet to China, chiefly to Peking; but Tibet does not 
produce any saffron, and imports it solely from Kashmir. Stuart 3 
says that "Ts'an hun hwa W, &H16 ('Red flower from Tsah,' that is, 
Central Tibet) is given by some foreign writers as another name for 
saffron, but this has not been found mentioned by any Chinese writer." 
In fact, that term is given in the Pen ts'ao kan mu H i* and the Ci wu 
mih H t'u k x ao of 1848, 6 where it is said to come from Tibet (Si-tsah) 
and to be the equivalent of the Fan hun hwa of the Pen ts'ao kan mu. 
Ts'an hwa is still a colloquial name for saffron in Peking; it is also called 
simply hun hwa ("red flower"). 6 By Tibetans in Peking I heard it 
designated gur-kum, ia-ka-ma, and dri-bzan ("of good fragrance"). 
Saffron is looked upon by the Chinese as the most valuable drug sent 
by Tibet, ts'an hiah ("Tibetan incense") ranking next. 

Li Si-cen 7 holds that there are two yil-kin ti£ #, — the yii-kin aromatic, 
the flowers of which only are used; and the yii-kin the root of which is 
employed. The former is the saffron {Crocus sativus); the latter, a 
Curcuma. As will be seen, however, there are at least three yii-kin. 

Of the genus Curcuma, there are several species in China and 
Indo-China, — C. leucorrhiza (yii-kin), C.longa (kian hwan §c or ^!iC, 

1 The Arabs first brought saffron to Spain; and from Arabic za'fardn are derived 
Spanish azafran, Portuguese agafrdo or azafrao, Indo-Portuguese safrdo, Italian 
zafferano, French safran, Rumanian sofrdn. The same Arabic root ('asfur, "yellow") 
has supplied also those Romance words that correspond to our safflow, safflower 
(Carthamus tinctorius), like Spanish azafranillo, alazor, Portuguese agafroa, Italian 
asforo, French safran; Old Armenian zavhran, New Armenian zafran; Russian 
safran; Uigur sakparan. 

2 Mat. m6d. et pharmacopee sino-annamites, p. 94. 

3 Chinese Materia Medica, p. 132. 

4 Ch. 4, p. 14 b. 

6 Ch. 4l p. 35 b. 

* It should be borne in mind that this name is merely a modern colloquialism, 
but hun hwa, when occurring in ancient texts, is not "saffron," but "safflower" 
(Carthamus tinctorius)', see below, p. 324. 

7 Pen ts'ao kan mu, Ch. 14, p. 18. 

Saffron and Turmeric 313 

"ginger-yellow"), C pallida, C. petiolata, C. zedoaria. Which particular 
species was anciently known in China, is difficult to decide; but it 
appears that at least one species was utilized in times of antiquity. 
Curcuma longa and C. leucorrhiza are described not earlier than theT'ang 
period, and the probability is that either they were introduced from the 
West; or, if on good botanical evidence it can be demonstrated that 
these species are autochthonous, 1 we are compelled to assume that 
superior cultivated varieties were imported in the T'ang era. In regard 
to yii-kin (C. leucorrhiza), Su Kun of the seventh century observes 
that it grows in Su (Se-c'wan) and Si-zun, and that the Hu call it 
$f ijt ma-lu, *mo-dzut (dzut), 2 while he states with reference to kian- 
hwah (C. longa) that the Zun 2$G A call it ^ Su, *dzut (dzut, dzur) ; 
he also insists on the close resemblance of the two species. Likewise 
C'en Ts'ah-k'i, who wrote in the first part of the eighth century, states 
concerning kian-hwan that the kind coming from the Western Bar- 
barians (Si Fan) is similar to yii-kin and Su yao £§ I3I. 3 Su Sun of the 
Sung remarks that yii-kin now occurs in all districts of Kwan-tuh and 
Kwah-si, but does not equal that of Se-c'wan, where it had previously 
existed. K'ou Tsun-si 4 states that yii-kin is not aromatic, and that in 
his time it was used for the dyeing of woman's clothes. Li Si-cen re- 
minds us of the fact that yii-kin was a product of the Hellenistic Orient 
(Ta Ts'in) : this is stated in the Wei lio of the third century, 5 and the 
Liah Su 6 enumerates yii-kin among the articles traded from Ta Ts'in 
to western India. 7 

The preceding observations, in connection with the foreign names 

1 According to Loureiro (Flora Cochin-Chinensis, p. 9), Curcuma longa 
grows wild in Indo-China. 

2 This foreign name has not been pointed out by Bretschneider or Stuart or 
any previous author. 

* This term is referred (whether correctly, I do not know) to Kcempferia 
pundurata (Stuart, Chinese Materia Medica, p. 227). Another name for this 
plant is J|r §£ jrj£ p'un-no Su (not mou), *buh-ha. Now, Ta Mih states that the 
Curcuma growing on Hai-nan is ^ ^ j^E P'un-no Su, while that growing in Kian-nan 
is kian-hwan (Curcuma longa). Kcempferia belongs to the same order as Curcuma, 
— Scitamineae. According to Ma Ci of the Sung, this plant grows in Si-zuh and in 
all districts of K wan-nan; it is poisonous, and the people of the West first test it 
on sheep : if these refuse to eat it, it is discarded. Chinese p'un-no, *buh-na, looks like 
a transcription of Tibetan bon-na, which, however, applies to aconite. 

* Pen ts'ao yen i, Ch. io, p. 3. 
6 San kwo U, Ch. 30, p. 13. 

6 Ch. 78, p. 7. 

7 The question whether in this case Curcuma or Crocus is meant, cannot be 
decided; both products were known in western Asia. C'en Ts'ah-k'i holds that the 
yii-kin of Ta Ts'in was safflower (see below). 

314 Sino-Iranica 

£m and ma-lu, are sufficient to raise serious doubts of the indigenous 
character of Curcuma; and for my part, I am strongly inclined to believe 
that at least two species of this genus were first introduced into Se-c'wan 
by way of Central Asia. This certainly would not exclude the possi- 
bility that other species of this genus, or even other varieties of the 
imported species, pre-existed in China long before that time; and this 
is even probable, in view of the fact that a fragrant plant yii Wt, which 
was mixed with sacrificial wine, is mentioned in the ancient Cou li, 
the State Ceremonial of the Cou Dynasty, and in the Li ki. The com- 
mentators, with a few exceptions, agree on the point that this ancient 
yii was a yii-kin; that is, a Curcuma. 1 

In India, Curcuma longa is extensively cultivated all over the coun- 
try, and probably so from ancient times. The plant (Sanskrit haridra) 
is already listed in the Bower Manuscript. From India the rhizome is 
exported to Tibet, where it is known as yun-ba or skyer-pa, the latter 
name originally applying to the barberry, the wood and root of which, 
like Curcuma, yield a yellow dye. 

Ibn al-Baitar understands by kurkum the genus Curcuma, not Cro- 
cus, as is obvious from his definition that it is the great species of the 
tinctorial roots. These roots come from India, being styled hard in 
Persian; this is derived from Sanskrit haridra (Curcuma longa). Ibn 
Hassan, however, observes that the people of Basra bestow on hard 
the name kurkum, which is the designation of saffron, and to which it 
is assimilated; but then he goes on to confound saffron with the root of 
wars, which is a Memecylon (see below). 2 Turmeric is called in Persian 
zird-lube or darzard ("yellow wood"). According to Garcia da Orta, 
it was much exported from India to Arabia and Persia; and there was 
unanimous opinion that it did not grow in Persia, Arabia, or Turkey, 
but that all comes from India. 3 

The name yii-kin, or with the addition hian ("aromatic"), 4 is fre- 
quently referred in ancient documents to two different plants of Indian 
and Iranian countries, — Memecylon tinctorium and Crocus sativus, the 

1 Cf. Bretschneider, Bot. Sin., pt. II, No. 408. 

2 Leclerc, Trait6 des simples, Vol. Ill, p. 167. 

3 C. Markham, Colloquies, p. 163. 

4 As a matter of principle, the term yii-kin hian strictly refers to saffron. It is 
this term which Bretschneider (Bot. Sin., pt. II, No. 408) was unable to identify, 
and of which Stuart (Chinese Materia Medica, p. 140) was compelled to admit, 
"The plant is not yet identified, but is probably not Curcuma." The latter remark 
is to the point. The descriptions we have of yii-kinhian, and which are given below, 
exclude any idea of a Curcuma. The modern Japanese botanists apply the term yii-kin 
hian (Japanese ukkonko) to Tulipa gesneriana, a flower of Japan (Matsumura, 
No. 3193)- 

Saffron and Turmeric 315 

latter possibly confounded again with Curcuma. 1 It is curious that 
in the entire Pen-is* ao literature the fact has been overlooked that under 
the same name there is also preserved the ancient description of a tree. 
This fact has escaped all European writers, with the sole exception of 
Palladius. In his admirable Chinese-Russian Dictionary 2 he gives 
the following explanation of the term yil-kin: "Designation of a tree 
in Ki-pin; yellow blossoms, which are gathered, and when they begin 
to wither, are pressed, the sap being mixed with other odorous sub- 
stances; it is found likewise in Ta Ts'in, the blossoms being like those 
of saffron, and is utilized in the coloration of wine." 

A description of this tree yii-kin is given in the Buddhist dictionary 
Yi ts'ie kin yin i 3 of a.d. 649 as follows: "This is the name of a tree, 
the habitat of which is in the country Ki-pin M SC (Kashmir) . Its 
flowers are of yellow color. The trees are planted from the flowers. 
One waits till they are faded; the sap is then pressed out of them and 
mixed with other substances. It serves as an aromatic. The grains 
of the flowers also are odoriferous, and are likewise employed as aro- 

I am inclined to identify this tree with Memecylon tinctorium, M. 
edule, or M. capitellatum (Melastomaceae), a very common, small tree 
or large shrub in the east and south of India, Ceylon, Tenasserim, and 
the Andamans. The leaves are employed in southern India for dyeing 
a "delicate yellow lake." The flowers produce an evanescent yellow. 4 
In restricting the habitat of the tree to Kashmir, Huan Yin is doubtless 
influenced by the notion that saffron (yil-kin) was an exclusive product 
of Kashmir (see below). 

The same tree is described by Abu Mansur under the name wars 
as a saffron-like plant of yellow color and fragrant, and employed by 
Arabic women for dyeing garments. 6 The ancients were not acquainted 

1 A third identification has been given by Bretschneider (Chinese Recorder, 
1871, p. 222), who thought that probably the sumbul (Sumbulus moschatus) is meant. 
This is a mistaken botanical name, but he evidently had in mind the so-called musk- 
root of Euryangium or Ferula sumbul, of musk-like odor and acrid taste. The only 
basis for this identification might be sought in the fact that one of the synonymes 
given for yii-kin hian in the Pen ts'ao is ts'ao Se hian ifL Jj$ ^ ("vegetable musk"); 
this name itself, however, is not explained. Saffron, of course, has no musk odor; 
and the term ts'ao Se hian surely does not relate to saffron, but is smuggled in here 
by mistake. The Tien hai yii hen U (Ch. 3, p. I b, see above, p. 228) also equates yii- 
kin hian with ts'ao Se hian, adding that the root is like ginger and colors wine yel- 
low. This would decidedly hint at a Curcuma. 

2 Vol. II, p. 202. 

3 Ch. 24, p. 8 (cf. Beginnings of Porcelain, p. 115; and above, p. 258). 

4 Watt, Dictionary of the Economic Products of India, Vol. V, p. 227. 
6 Achundow, Abu Mansur, p. 145. 

316 Sino-Iranica 

with this dye. Abu Hanifa has a long discourse on it. 1 Ibn Hassan 
knew the root of wars, and confounded it with saffron. 2 Ibn al-Baitar 
offers a lengthy notice of it. 3 Two species are distinguished, — one from 
Ethiopia, black, and of inferior quality; and another from India, of a 
brilliant red, yielding a dye of a pure yellow. A variety called bdrida 
dyes red. It is cultivated in Yemen. Also the association with Cur- 
cuma and Crocus is indicated. Isak Ibn Amran remarks, "It is said 
that wars represents roots of Curcuma, which come from China and 
Yemen"; and Ibn Massa el-Basri says, "It is a substance of a brilliant 
red which resembles pounded saffron." This explains why the Chinese 
included it in the term yii-kin. Leclerc also has identified the wars 
of the Arabs with Memecylon tinctorium, and adds, "L'ouars n'est pas 
le produit exclusif de l'Arabie. On le rencontre abondamment dans 
l'Inde, notamment aux environs de Pondich6ry qui en a envoye" en 
Europe, aux dernieres expositions. II s'appelle kana dans le pays." 4 
The Yamato honzo speaks of yii-kin as a dye-stuff coming from Siam; 
this seems to be also Memecylon. 

The fact that the Chinese included the product of Memecylon in 
the term yii-kin appears to indicate that this cheap coloring-matter 
was substituted in trade for the precious saffron. 

While the Chinese writers on botany and pharmacology have over- 
looked yii-kin as the name of a tree, they have clearly recognized that 
the term principally serves for the designation of the saffron, the product 
of the Crocus sativus. This fact is well borne out by the descriptions 
and names of the plant, as well as by other evidence. 

The account given of Central India in the Annals of the Liang 
Dynasty 5 expressly states that yii-kin is produced solely in Kashmir 
(Ki-pin), that its flower is perfectly yellow and fine, resembling the 
flower fu-yun (Hibiscus mutabilis). Kashmir was always the classical 
land famed for the cultivation of saffron, which was (and is) thence 
exported to India, Tibet, Mongolia, and China. In Kashmir, Uddiyana, 

1 Achundow, Abu Mansur, p. 272. 

2 Leclerc, Trait6 des simples, Vol. Ill, p. 167. 
8 Ibid., p. 409. 

4 Arabic wars has also been identified with Flemingia congesta (Watt, Diction- 
ary, Vol. Ill, p. 400) and Mallotus philippinensis (ibid., Vol. V, p. 114). The whole 
subject is much confused, particularly by Fluckiger and Hanbury (Pharma- 
cographia, p. 573; cf. also G. Jacob, Beduinenleben, p. 15, and Arab. Geographen, 
p. 166), but this is not the place to discuss it. The Chinese description of the yii-kin 
tree does not correspond to any of these plants. 

6 Lian Su, Ch. 54, p. 7 b. This work was compiled by Yao Se-lien in the first 
half of the seventh century from documents of the Liang dynasty, which ruled from 

A.D. 502 to 556. 

Saffron and Turmeric 317 

and Jagu4a (Zabulistan) it was observed by the famous pilgrim Huan 
Tsan in the seventh century. 1 The Buddhist traveller Yi Tsin (671-695) 
attributes it to northern India. 2 

The earliest description of the plant is preserved in the Nan lou i 
wu ci, written by Wan Cen in the third century a.d., 8 who says, "The 
habitat of yii-kin is in the country Ki-pin (Kashmir), where it is culti- 
vated by men, first of all, for the purpose of being offered to the Buddha. 
After a few days the flower fades away, and then it is utilized on 
account of its color, which is uniformly yellow. It resembles the fu-yun 
(Hibiscus) and a young lotus (Nelumbium speciosum), and can render 
wine aromatic." This characteristic is fairly correct, and unequivocally 
applies to the Crocus, which indeed has the appearance of a liliaceous 
plant, and therefore belongs to the family Irideae and to the order 
Liliiflorae. The observation in regard to the short duration of the 
flowers is to the point. 4 

In a.d. 647 the country Kia-p'i f&fl lit in India offered to the Court 
yii-kin hian, which is described on this occasion as follows: "Its leaves 
are like those of the mai-men-tun ££ ffi %- (Ophiopogon spicatus). It 
blooms in the ninth month. In appearance it is similar to fu-yuh 
(Hibiscus mutabilis). It is purple-blue 3£ i§ in color. Its odor may be 
perceived at a distance of several tens of paces. It flowers, but 
does not bear fruit. In order to propagate it, the root must be 
taken." 5 

1 S. Julien, Memoires sur les contrdes occidentales, Vol. I, pp. 40, 131; Vol. 
II, p. 187 (story of the Saffron-StQpa, ibid., Vol. I, p. 474; or S. Beal, Buddhist 
Records, Vol. II, p. 125); W. W. Rockhill, Life of the Buddha, p. 169; S. Levi, 
Journal asiatique, 1915, I, pp. 83-85. 

1 Takakusu's translation, p. 128; he adds erroneously, "species of Curcuma." 

8 Pen ts'ao kan mu, Ch. 14, p. 22. 

4 Compare Pliny's (xxi, 17, §34) description of Crocus: "Floret vergiliarum 
occasu paucis diebus folioque florem expellit. Viret bruma et colligitur; siccatur 
umbra, melius etiam hiberna." 

5 T'an hui yao, Ch. 200, pp. 14 a-b. This text was adopted by the Pen Is'ao 
kan mu (Ch. 14, p. 22), which quotes it from the T'ang Annals. Li §i-cen comments 
that this description agrees with that of the Nan lou i wu U, except in the colors of 
the flower, which may be explained by assuming that there are several varieties; in 
this he is quite correct. The flower, indeed, occurs in a great variation of colors, — 
purple, yellow, white, and others. W. Woodville (Medical Botany, Vol. IV, p. 763) 
gives the following description of Crocus: "The root is bulbous, perennial: the flower 
appears after the leaves, rising very little above the ground upon a slender succulent 
tube: the leaves rise higher than the flower, are linear, simple, radical, of a rich 
green colour, with a white line running in the centre, and all at the base inclosed 
along with the tube of the flower in a membranous sheath. The flower is large, of a 
bluish purple, or lilac colour: the corolla consists of six petals, which are nearly 
elliptical, equal, and turned inwards at the edges. The filaments are three, short, 
tapering, and support long erect yellow antherae. The germen is roundish, from 

318 Sino-Iranica 

The last clause means that the plant i,s propagated from 
bulbs. There is a much earlier tribute-gift of saffron on record. In 
a.d. 519, King Jayavarman of Fu-nan (Camboja) offered saffron with 
storax and other aromatics to the Chinese Court. 1 Accordingly we have 
to assume that in the sixth century saffron was traded from India to 
Camboja. In fact we know from the T'ang Annals that India, in her 
trade with Camboja and the anterior Orient, exported to these coun- 
tries diamonds, sandal-wood, and saffron. 2 The T'ang Annals, further, 
mention saffron as a product of India, Kashmir, Uddiyana, Jagucja, 
and Baltistan. 3 In a.d. 719 the king of Nan (Bukhara) presented 
thirty pounds of saffron to the Chinese Emperor. 4 

Li Si-cen has added to his notice of yii-kin hian a Sanskrit name 
^ $£ m Va-kii-mo, *dza-gu-ma, which he reveals from the Suvar- 
naprabhasa-sutra. 5 This term is likewise given, with the translation 
yii-kin, in the Chinese-Sanskrit Dictionary Fan yi min yi tsi. 6 This name 
has been discussed by me and identified with Sanskrit jaguda through 
the medium of a vernacular form *jaguma, the ending -ma corresponding 
to that of Tibetan la-ka-ma? 

A singular position is taken by C'en Ts'ah-k'i, who reports, " Yii-kin 
aromatic grows in the country Ta Ts'in. It flowers in the second or 
third month, and has the appearance of the hun-lan (safflower, Car- 
thamus tinctorius) . 8 In the fourth or fifth month the flowers are gathered 
and make an aromatic." This, of course, cannot refer to the saffron 
which blooms in September or October. C'en Ts'an-k'i has created 
confusion, and has led astray Li Si-cen, who wrongly enumerates hun- 
lan hwa among the synonymes of yii-kin hian. 

The inhabitants of Ku-lin (Quilon) W» ^ rubbed their bodies with 

which issues a slender style, terminated by three long convoluted stigmata, of a 
deep yellow colour. The capsule is roundish, three-lobed, three-celled, three-valved, 
and contains several round seeds. It flowers in September and October." 

1 According to the Lian $u; cf. Pelhot, Bull, de V Ecole jrancaise, Vol. Ill, p. 270. 

2 T'an su, Ch. 221 A, p. 10 b. 

3 Kiu T'an su, Ch. 221 B, p. 6; 198, pp. 8 b, 9; T'an su, Ch. 221 A, p. 10 b; cf. 
Chavannes (Documents sur les Tou-kiue occidentaux, pp. 128, 150, 160, 166), 
whose identification with Curcuma longa is not correct. 

4 Chavannes, ibid., p. 203. 

5 The passage in which Li §i-6en cites this term demonstrates clearly that he 
discriminated well between Crocus and Curcuma; for he adds that " c'a-ku-mo is 
the aromatic of the yii-kin flower (Crocus), but that, while it is identical in name 
with the yii-kin root (Curcuma) utilized at the present time, the two plants are 

6 Ch. 8, p. 10 b. 

7 T'oung Pao, 1916, p. 458. 

8 See below, p. 324. 

Saffron and Turmeric 319 

yii-kin after every bath, with the intention of making it resemble the 
"gold body" of a Buddha. 1 Certainly they did not smear their bodies 
with "turmeric," 2 which is used only as a dye-stuff, but with saffron. 
Annamese mothers rub the bodies of their infants with saffron-powder 
as a tonic to their skin. 3 

The Ain-i Akbari, written 1597 in Persian by Abul Fazl 'Allami 
(1 551-1602), gives detailed information on the saffron cultivation in 
Kashmir, 4 from which the following extract may be quoted: "In the 
village of Pampur, one of the dependencies of Vihl (in Kashmir) , there 
are fields of saffron to the extent of ten or twelve thousand bighas, a 
sight that would enchant the most fastidious. At the close of the 
month of March and during all April, which is the season of cultivation, 
the land is plowed up and rendered soft, and each portion is prepared 
with the spade for planting, and the saffron bulbs are hard in the ground. 
In a month's time they sprout, and at the close of September, it is at 
its full growth, shooting up somewhat over a span. The stalk is white, 
and when it has sprouted to the height of a finger, one bud after another 
begins to flower till there are eight flowers. It has six lilac-tinted petals. 
Usually among six filaments, three are yellow and three ruddy. The 
last three yield the saffron. [There are three stamens and three stigmas 
in each flower, the latter yielding the saffron.] When the flowers are 
past, leaves appear upon the stalk. Once planted it will flower for six 
years in succession. The first year, the yield is small : in the second as 
thirty to ten. In the third year it reaches its highest point, and the 
bulbs are dug up. If left in the same soil, they gradually deteriorate, 
but if taken up, they may be profitably transplanted." 

The Emperor Jahangir was deeply impressed by the saffron planta- 
tions of Kashmir, and left the following notes in his Memoirs: 6 — 

"As the saffron was in blossom, his Majesty left the city to go to 
Pampur, which is the only place in Kashmir where it flourishes. Every 
parterre, every field, was, as far as the eye could reach, covered with 
flowers. The stem inclines toward the ground. The flower has five 
petals of a violet color, and three stigmas producing saffron are found 
within it, and that is the purest saffron. In an ordinary year, 400 

1 Lin wai tai ta, Ch. 2, p. 13. 

2 Hirth, Chau Ju-kua, p. 91. 

8 Perrot and Hurrier, Mat. m6d. et pharmacop£e sino-annamites, p. 94. 
Cf. also Marco Polo's observation (Yule's edition, Vol. II, p. 286) that the faces 
of stuffed monkeys on Java are daubed with saffron, in order to give them a manlike 

4 Translation of H. Blochmann, Vol. I, p. 84; Vol. II, p. 357. 

6 H. M. Eluot, History of India as told by Its Own Historians, Vol. VI, p. 375 

320 Sino-Iranica 

maunds, or 3200 Khurasanl maunds, are produced. Half belongs to 
the Government, half to the cultivators, and a sir sells for ten rupees; 
but the price sometimes varies a little. It is the established custom to 
weigh the flowers, and give them to the manufacturers, who take them 
home and extract the saffron from them, and upon giving the extract, 
which amounts to about one-fourth weight of the flower, to the public 
officers, they receive in return an equal weight of salt, in lieu of money 

The ancient Chinese attribute saffron not only to Kashmir, but also 
to Sasanian Persia. The Lou Su 1 enumerates yti-kin among the products 
of Po-se (Persia) ; so does the Sui $u? In fact, Crocus occurs in Persia 
spontaneously, and its cultivation must date from an early period. 
Aeschylus alludes to the saffron-yellow footgear of King Darius. 3 
Saffron is mentioned in Pahlavi literature (above, p. 193). The plant is 
well attested for Derbend, Ispahan, and Transoxania in the tenth 
century by Istaxri and Edrisi. 4 Yaqut mentions saffron as the principal 
production of Rud-Derawer in the province Jebal, the ancient Media, 
whence it was largely exported. 6 Abu Mansur describes it under the 
Arabic name zafardn. 6 The Armenian consumers esteem most highly 
the saffron of Khorasan, which, however, is marketed in such small 
quantities that the Persians themselves must fill the demand with 
exportations from the Caucasus. 7 According to Schlimmer, 8 part of 
the Persian saffron comes from Baku in Russia, another part is culti- 
vated in Persia in the district of Kain, but in quantity insufficient to 
fill the demand. In two places, — Rudzabar (identical with the above 
Rud-Derawer), a mountainous tract near Hamadan, and Mount 
Derbend, where saffron cultivation had been indicated by previous 
writers, — he was unable to find a trace of it. 

It is most probable that it was from Persia that the saffron-plant 
was propagated to Kashmir. A reminiscence of this event is preserved 
in the Sanskrit term vdhlika, a synonyme of "saffron," which means 
"originating from the Pahlava." 9 The Buddhists have a legend to the 

1 Ch. 50, p. 6. 

2 Ch. 83, p. 7 b; also Wei $u, Ch. 102, p. 5 b. 
8 Hehn, Kulturpflanzen, p. 264. 

4 A. Jaubert, Geographie, pp. 168, 192. 

6 B. de Meynard, Dictionnaire geogr. de la Perse, p. 267. See also G. Fer- 
rand, Textes relatifs a 1'ExtrSme-Orient, Vol. II, pp. 61 8, 622. 

6 Achundow, Abu Mansur, p. 76. 

7 E. Seidel, Mechithar, p. 151. Chardin (Voyages en Perse, Vol. II, p. 14) 
even says that the saffron of Persia is the best of the world. 

8 Terminologie, p. 165. 

9 Cf. T'oung Pao, 1916, p. 459. 

Saffron and Turmeric 321 

effect that Madhyantika, the first apostle of Buddha's word in Kashmir, 
planted the saffron there. 1 If nothing else, this shows at least that the 
plant was regarded as an introduction. The share of the Persians in the 
distribution of the product is vividly demonstrated by the Tibetan 
word for "saffron, "kur-kum, gur-kum, gur-gum, whichis directly traceable 
to Persian kurkum or karkatn, but not to Sanskrit kunkuma. 2 The 
Tibetans carried the word to Mongolia, and it is still heard among the 
Kalmuk on the Wolga. By some, the Persian word (Pahlavi kulkem) 
is traced to Semitic, Assyrian karkutna, Hebrew karkom, Arabic kurkum; 
while others regard the Semitic origin as doubtful. 3 It is beyond the 
scope of this notice to deal with the history of saffron in the west and 
Europe, on which so much has been written. 4 

From the preceding investigation it follows that the word yii-kin 
& 4£, owing to its multiplicity of meaning, offers some difficulty to 
the translator of Chinese texts. The general rule may be laid down that 
yii-kin, whenever it hints at a plant or product of China, denotes a 
species of Curcuma, but that, when used with reference to India, Indo- 
China, and Iran, the greater probability is in favor of Crocus. The term 
yii-kin hian ("yu-kin aromatic"), with reference to foreign countries, 
almost invariably appears to refer to the latter plant, which indeed 
served as an aromatic; while the same term, as will be seen below, with 
reference to China, again denotes Curcuma. The question may now be 
raised, What is the origin of the word yii-kin? And what was its original 
meaning? In 1886 Hirth 5 identified yii-kin with Persian karkam 
("saffron"), and restated this opinion in 1911, 6 by falling back on an 
ancient pronunciation *hat-kam. Phonetically this is not very con- 
vincing, as the Chinese would hardly have employed an initial h for 

1 Schiefner, Taranatha, p. 13; cf. also J. Przyluski, Journal asiatique, 19 14 
II. P- 537- 

* Voung Pao, 1916, p. 474. Cf. also Sogdian kurkumba and Tokharian kurkama. 

8 Horn, Grundriss der iranischen Philologie, Vol. I, pt. 2, p. 6. Besides kurkum, 
there are Persian kdkbdn and kdftfa, which denote "saffron in the flower." Old 
Armenian k'rk'um is regarded as a loan from Syriac kurkema (H^bschmann, Armen. 
Gram., p. 320). 

* In regard to saffron among the Arabs, see Leclerc, Traite" des simples, 
Vol. II, pp. 208-210. In general cf. J. Beckmann, Beytrage zur Geschichte der 
Erfindungen, 1784, Vol. II, pp. 70-91 (also in English translation); FlCckiger and 
Hanbury, Pharmacographia, pp. 663-669; A. de Candolle, Geographie botanique, 
p. 857, and Origin of Cultivated Plants, p. 166; Hehn, Kulturpflanzen (8th ed.), 
pp. 264-270; Watt, Dictionary, Vol. II, p. 592; W. Heyd, Histoire du commerce du 
levant, Vol. II, p. 668, etc. 

6 Journal China Branch Roy. As. Soc, Vol. XXI, p. 221. 

* Chau Ju-kua, p. 91. . 

322 Sino-Iranica 

the reproduction of a foreign k; but the character yu in transcriptions 
usually answers to *ut, ud. The whole theory, however, is exposed to 
much graver objections. The Chinese themselves do not admit that 
yii-kin represents a foreign word; nowhere do they say that yii-kin is 
Persian, Sanskrit, or anything of the sort; on the contrary, they regard 
it as an element of their own language. Moreover, if yii-kin should 
originally designate the saffron, how, then, did it happen that this alleged 
Persian word was transferred to the genus Curcuma, some species of 
which are even indigenous to China, and which, at any rate, has been 
acclimated there for a long period? The case, indeed, is not simple, and 
requires closer study. Let us see what the Chinese have to say con- 
cerning the word yii-kin. Pelliot 1 has already clearly, though briefly, 
outlined the general situation by calling attention to the fact that as 
early as the beginning of the second century, yii-kin is mentioned in 
the dictionary Swo wen as the name of an odoriferous plant, offered as 
tribute by the people of Yu, the present Yu-lin in Kwan-si Province; 
hence he inferred that the sense of the word should be "gold of Yu," 
in allusion to the yellow color of the product. We read in the Swi kin 
lu *K % W as follows: "The district Kwei-lin & # M of the Ts'in 
dynasty had its name changed into the Yu-lin district ^ ^ fP in the 
sixth year of the period Yuan-tin (in b.c.) of the Emperor Wu of the 
Han dynasty. Wan Mail made it into the Yu-p'ih district Wt Z P. Yin 
Sao J® S{5 [second century a.d.], in his work Ti li fun su ki ife S M, 
f&wE, says, 'The Cou li speaks of the yii Zen&K ('officials in charge of 
the plant yu'), who have charge of the jars serving for libations; when- 
ever libations are necessary for sacrifices or for the reception of guests, 
they attend to the blending of the plant yii with the odoriferous wine 
l K an, pour it into the sacred vases, and arrange them in their place.' 3 
Yii is a fragrant plant. Flowers of manifold plants are boiled and mixed 
with wine fermented by means of black millet as an offering to the 
spirits: this is regarded by some as what is now called yii-kin hian 
W 4£ # (Curcuma) ; while others contend that it was brought as 
tribute by the people of Yu, thus connecting the name of the plant 
with that of the clan and district of Yii." The latter is the explanation 

1 Bull, de I'Ecole frangaise, Vol. Ill, p. 270. 

2 This work is a commentary to the Swi kin, a canonical book on water-courses, 
supposed to have been written by San K'in under the Later Han dynasty, but it 
was elaborated rather in the third century. The commentary is due to Li Tao-yuan 
of the Hou Wei period, who died in A.D. 527 (his biography is in Wei $u, Ch. 89; 
Pei U, Ch. 27). Regarding the various editions of the work, see Pelliot, Bull, de 
I'Ecole frangaise, Vol. VI, p. 364, note 4. 

s Cf. Biot, Le Tcheou-li, Vol. I, p. 465. 

Saffron and Turmeric 323 

favored by the Swo wen. 1 Both explanations are reasonable, but only 
one of the two can be correct. 2 My own opinion is this: yii is an ancient 
Chinese name for an indigenous Chinese aromatic plant; whether 
Curcuma or another genus, can no longer be decided with certainty. 3 
The term yii-kin means literally "gold of the yii plant," "gold" re- 
ferring to the yellow rhizome, 4 yii to the total plant-character; the con- 
crete significance, accordingly, is "jw-rhizome" or "yw-root." I do not 
believe, however, that yii-kin is derived from the district or clan of Yu; 
for this is impossible to assume, since yii as the name of a plant existed 
prior to the name of that district. This is clearly evidenced by the 
text of the Swi kin lu: for it was only in in B.C. that the name Yu-lin 
("Grove of the Yii Plant") came into existence, being then substituted 
for the earlier Kwei-lin ("Grove of Cinnamomum cassia"). It is the 
plant, consequently, which lent its name to the district, not the dis- 
trict which named the plant. As in so many cases, the Chinese con- 
found cause and effect. The reason why the name of this district was 
altered into Yu-lin is now also obvious. It must have been renowned 
under the Han for the wealth of its yii-kin plants, which was less con- 
spicuous under the Ts'in, when the cassia predominated there. At 
any rate, yii-kin is a perfectly authentic and legitimate constituent 
of the Chinese language, and not a foreign word. It denotes an indig- 
enous Curcuma; while under the T'ang, as we have seen, additional 
species of this genus may have been introduced from abroad. The word 
yii-kin then underwent a psychological treatment similar to yen-ci: 
as yen-fi, "safflower," was transformed to any cosmetic or rouge, so yii-kin 
"turmeric," was grafted on any dyes producing similar tinges of yellow. 
Thus it was applied to the saffron of Kashmir and Persia. 

1 The early edition of this work did not contain the form yii-kin, but merely the 
plain, ancient yii. Solely the Fan yi min yi tsi (Ch. 8, p. 10 b) attributes ( I believe, 
erroneously) the term yii-kin to the Swo wen. 

2 Li Si-cen says that the district Yu-lin of the Han period comprises the territory 
of the present cou >)]\ of Sun ?|f[, Liu $P, Yuh f , and Pin ^ of Kwan-si and Kwei- 
cou, and that, according to the Ta Min i t'un li, only the district of Lo-S'en ff; fjjfc 
in Liu-cou fu (Kwan-si) produces yii-kin hiah, which is that here spoken of (that is, 
Crocus), while in fact Curcuma must be understood. 

* There is also the opinion that the ancient yii must be a plant similar to Ian 
|5ij, an orchidaceous plant (see the P'i ya of Lu Tien and the T'un ci of Cen Tsiao). 

* Pallegoix (Description du royaume Thai ou Siam, Vol. I, p. 126) says, "Le 
curcuma est une racine bulbeuse et charnue, d'un beau jaune d'or." 


17. A. de Candolle, 1 while maintaining that the cultivation of 
safflower 2 (Carthamus tinctorius) is of ancient date both in Egypt and 
India, asserts on Bretschneider's authority that the Chinese received it 
only in the second century B.C., when Can K'ien brought it back from 
Bactriana. The same myth is repeated by Stuart. 3 The biography 
of the general and the Han Annals contain nothing to this effect. Only 
the Po wu li enumerates hwan Ian ]H ^ in its series of Can-K'ien plants, 
adding that it can be used as a cosmetic (yen-li %& ^). 4 The Ku kin 
lu, while admitting the introduction of the plant from the West, makes 
no reference to the General. The TsH min yao Su discusses the method 
of cultivating the flower, but is silent as to its introduction. The fact 
of this introduction cannot be doubted, but it is hardly older than the 
third or fourth century a.d. under the Tsin dynasty. The introduction 
of safflower drew the attention of the Chinese to an indigenous wild 
plant (Basella rubra) which yielded a similar dye and cosmetic, and 
both plants and their products were combined or confounded under 
the common name yen-li. 

Basella rubra, a climbing plant of the family Basellaceae, is largely 
cultivated in China (as well as in India) on account of its berries, which 
contain a red juice used as a rouge by women and as a purple dye for 
making seal-impressions. This dye was the prerogative of the highest 

1 Origin of Cultivated Plants, p. 164. 

2 Regarding the history of this word, see Yule, Hobson-Jobson, p. 779. 

8 Chinese Materia Medica, p. 94. It is likewise an erroneous statement of Stuart 
that Tibet was regarded by the Chinese as the natural habitat of this plant. This is 
due to a confusion with the term Si-ts'an hun hwa (" red flower of Tibet "), which refers 
to the saffron, and is so called because in modern times saffron is imported into 
China from Kashmir by way of Tibet (see p. 312). Neither Carthamus nor saffron is 
grown in the latter country. 

4 Some editions of the Po wu li add, "At present it has also been planted in 
the land of Wei ffi (China)," which might convey the impression that it had only 
been introduced during the third century a.d., the lifetime of Can Hwa, author of 
that work. In the commentary to the Pei hu lu (Ch. 3, p. 12), the Po wu li is quoted 
as saying, "The safflower {hun hwa ;££ /f6> 'red flower') has its habitat in Persia, 
Su-le (Kashgar), and Ho-lu $f jjifc. Now that of Liah-han U£ (H is of prime quality, 
a tribute of twenty thousand catties being annually sent to the Bureau of Weaving 
and Dyeing." The term hun hwa in the written language does not refer to "saffron," 
but to "safflower." Java produced the latter (Javanese kasumba), not saffron, as 
translated by Hirth (Chau Ju-kua, p. 78). The Can-K'ien story is repeated in the 
Hwa kin of 1688 (Ch. 5, p. 24 b). 


Safflower 325 

boards of the capital, the prefects of Sun-t'ien and Mukden, and all 
provincial governors. 1 Under the name lo k'wei $& ^ it is mentioned 
by T'ao Hun-kin (a.d. 451-536), who refers to its cultivation, to the 
employment of the leaves as a condiment, and to the use of the berries 
as a cosmetic. 2 This probably came into use after the introduction of 
safflower. The Ku kin lu* written by Ts'ui Pao in the middle of the 
fourth century, states, "The leaves of yen-li fS ;£ resemble those of 
the thistle (ki X5) and the p'u-kun ffl & (Taraxacum officinalis). Its 
habitat is in the Western Countries ® 32f , where the natives avail them- 
selves of the plant for dyeing, and designate it yen-li & ;£, while the 
Chinese call it hun-lan (%£. IE 'red indigo/ Carthamus tinctorius) ; 
and the powder obtained from it, and used for painting the face, is 
styled yen-li Jen $h [At present, because people value a deep-red 
color $$, they speak of the yen-li flower which dyes; the yen-li flower, 
however, is not the dye-plant yen-li, but has its own name, hun-lan 
(Carthamus tinctorius). Of old, the color intermediate between IH f$ 
and white is termed hun ifct, and this is what is now styled hun-lan.]" 4 
It would follow from this text that Basella was at an early date con- 
founded with Carthamus, but that originally the term yen-li related to 
Carthamus only. 

The Pei hu lu 5 contains the following information in regard to the 
yen-li flower: "There is a wild flower growing abundantly in the 
rugged mountains of Twan-cou m 'Hi. 6 Its leaves resemble those of the 
Ian ^ (Indigofera) ; its flowers, those of the liao H (Polygonum, prob- 
ably P. tinctorium). The blossoms Ht, when pulled out, are from two 
to three inches long, and yield a green-white pigment. It blooms in 
the first month. The natives gather the bursting seeds while still in 
their shells, in order to sell them. They are utilized in the preparation 
of a cosmetic %& ;£ %&, and particularly also for dyeing pongee and 
other silks. Its red is not inferior to that of the Ian flower. Si Ts'o-c'i 

1 P. Hoang, Melanges stir l'administration, pp. 80-81. 

* Bretschneider, Bot. Sin., pt. II, No. 148; pt. Ill, No. 258. 

3 Ch. c, p. 5 (ed. of Han Wei ts'un Su). In regard to the historicity of this work, 
the critical remarks of the Imperial Catalogue (cf. Wylie, Notes on Chinese Litera- 
ture, p. 159) must be kept in mind. Cf. also above, p. 242. 

4 The passage enclosed in brackets, though now incorporated in the text of the 
Ku kin lu, is without any doubt later commentatorial wisdom. This is formally 
corroborated by the Pei hu lu (Ch. 3, p. 12), which omits all this in quoting the 
relevant text of the Ku kin lu. 

6 Ch. 3, p. 11 (see above, p. 268). 

6 Name of the prefecture of Cao-k'in Ql fg£ in Kwan-tun Province. This 
wild flower is Basella rubra. 

326 Sino-Iranica 

f? J£ $t, in his Yii sie H lun hi % $t f# *f* it, says, 1 'These are hun- 
lan (Carthamus) : 2 did you know these previously, Sir, or not? The 
people of the north gather these flowers, and dye materials a red-yellow 
by rubbing their surface with it. The fresh blossoms are made into a 
cosmetic. 3 Women, when dressing, use this pigment, it being the fashion 
to apply only a piece the size of a small bean. When distributed evenly, 
the paint is pleasing, as long as it is fresh. In my youth I observed this 
cosmetic again and again; and to-day I have for the first time beheld 
the hun-lan flower. Afterwards I shall raise its seeds for your benefit, 
Sir. The Hiun-nu styled a wife yen-li f8 J£, 4 a word just as pleasing as 
yen-U ffl ;£ ('cosmetic ') . The characters P! and #9 have the same sound 
yen; the character J£ has the sound ;£ li. I expect you knew this 
before, Sir, or you may read it up in the Han Annals.' Ceh K'ien H5 ^ 5 
says that a cosmetic may be prepared from pomegranate flowers." 6 

The curious word yen-li has stirred the imagination of Chinese 
scholars. It is not only correlated with the Hiuh-nu word yen-U, as 
was first proposed by Si Ts'o-S'i, but is also connected with a Yen-ci 
mountain. Lo Yuan, in his Er ya i, remarks that the Hiun-nu had a 
Yen-ci mountain, and goes on to cite a song from the Si ho kiu Si M $f 
H ^, 7 which says, "If we lose our K'i-lien mountain tf(5 51 ill, 8 we cause 
our herds to diminish in number; if we lose our Yen-ci mountain, we 
cause our women to go without paint." 9 The Pet pien pei tui At jft 
4if Ws, a work of the Sung period, states, "The yen-li M 3£ of the Yen-ci 
mountain M ;£ \U is the yen-li ^ flit of the present time. This moun- 

1 This author is stated to have lived under the Tsin dynasty (a.d. 265-419) 
in the T'u I« tsi Ven, XX, Ch. 158, where this passage is quoted; but his book is 
there entitled Yii yen wan $u |^ ^ 3E lif- The same passage is inserted in the 
Er ya i of Lo Yuan ft I$f of the twelfth century, where the title is identical with 
that given above. V 

* In the text of the T'u Su: "At the foot of the mountain there are hurt Ian." 
8 Carthamus was already employed for the same purposes in ancient Egypt. 

* This is the Hiun-nu word for a royal consort, handed down in the Han Annals 
{TsHen Han Isu, Ch. 94 A, p. 5). See my Language of the Yue-chi, p. 10. 

6 Author of the lost Hu pen ts'ao (above, p. 268). 

6 Then follow a valueless anecdote anent a princess of the T'ang dynasty pre- 
paring a cosmetic, and the passage of the Ku kin iu given above. 

7 Mentioned in the T'ang literature, but seems to date from an earlier period 
(Bretschneider, Bot. Sin., pt. 1, p. 190). 

8 A mountain-range south-west of Kan Sou in Kan-su (Si ki, Ch. 123, p. 4). 
The word k'i-lien belongs to the language of the Hiun-nu and means "heaven." 
In my opinion, it is related to Manchu kulun, which has the same meaning. The 
interpretations given by Watters (Essays, p. 362) and Shiratori (Sprache der 
Hiung-nu, p. 8) are not correct. 

* The same text is quoted in the commentary to the Pei hu lu (Ch. 3, p. 11 b). 

Safflower 327 

tain produces hun-lan (Carthamus) which yields yen-U ('cosmetic')." 
All this, of course, is pure fantasy inspired by the homophony of the two 
words yen-U ("cosmetic") and Hiun-nu yen-U ("royal consort"). 
Another etymology propounded by Fu Hou ffc §1 in his Cun hwa ku 
kin lu r f 1 ii"fi ,> T^ (tenth century) is no more fortunate: he explains 
that yen-U is produced in the country Yen #5, and is hence styled fflk Ha 
yen-U ("sap of Yen"). Yen was one of the small feudal states at the 
time of the Cou dynasty. This is likewise a philological afterthought, 
for there is no ancient historical record to the effect that the state of 
Yen should have produced (exclusively or pre-eminently) Basella or 
Carthamus. It is perfectly certain that yen-U is not Chinese, but the 
transcription of a foreign word: this appears clearly from the ancient 
form %& ;£, which yields no meaning whatever; ;£, as is well known, 
being a favorite character in the rendering of foreign words. This is 
further corroborated by the vacillating modes of writing the word, 
to which Li Si-cen adds J^I %&, 1 while he rejects as erroneous ^ Jl£ 
and JJB ^, and justly so. Unfortunately we are not informed as to the 
country or language from which the word was adopted: the Ku kin 
£u avails itself only of the vague term Si fan ("Western Countries"), 
where Carthamus was called yen-U; but in no language known to me is 
there any such name for the designation of this plant or its product. 
The Sanskrit name for safflower is kusumbha; and if the plant had come 
from India, Chinese writers would certainly not have failed to express 
this clearly. The supposition therefore remains that it was introduced 
from some Iranian region, and that yen-U represents a word from an 
old Iranian dialect now extinct, or an Iranian word somehow still 
unknown. The New-Persian name for the plant is gawdlila; in Arabic 
it is qurtum. 2 

Li Si-cen distinguishes four kinds of yen-U: (1) From Carthamus 
tinctorius, the juice of the flowers of which is made into a rouge (the 
information is chiefly drawn from the Ku kin £u, as cited above). 
(2) From Basella rubra, as described in the Pei hu lu. (3) From the 
ian-liu tfi fi§ flower [unidentified, perhaps a wild pomegranate : above, 
p. 281], described in the Hu pen ts'ao. (4) From the tree producing 
gum lac (tse-kun ^ UJr 1 ) , 8 this product being styled 1&8 %& Ba hu yen-U 
("foreign cosmetic") and described in the Nan hai yao p'u $j $£ & M 
of Li Sun 9 *%.* "At present," Li Si-cen continues, "the southerners 

1 Formed with the classifier 155, "red." 

2 Achundow, Abu Mansur, p. 105. 
1 See below, p. 476. 

4 He lived in the second half of the eighth century. 

328 Sino-Iranica 

make abundant use of tse-kun cosmetic, which is commonly called 
tse-kuh. In general, all these substances may be used as remedies in 
blood diseases. 1 Also the juice from the seeds of lo k'wei $£ f£ (Basella 
rubra) may be taken, and, mixed evenly with powder, may be applied 
to the face. Also this is styled hu yen-Zi." Now it becomes clear why 
Basella rubra, a plant indigenous to China, is termed hu yen-Zi in the 
T'un Zi of Ceh Tsiao and by Ma Ci of the tenth century: this name 
originally referred to the cosmetic furnished by Butea frondosa or other 
trees on which the lac-insect lives, 2 — trees growing in Indo-China, the 
Archipelago, and India. This product, accordingly, was foreign, and 
hence styled "foreign cosmetic" or "cosmetic of the barbarians" 
(hu yen-Zi). Since Basella was used in the same manner, that name 
was ultimately transferred also to the cosmetic furnished by this 
indigenous plant. 

What is not stated by Li Si-cen is that yen-Zi is also used with 
reference to Mirabilis jalapa, because from the flowers of this plant is 
derived a red coloring-matter often substituted for carthamine. 3 It 
is obvious that the term yen-li has no botanical value, and for many 
centuries has simply had the meaning "cosmetic." 

Fan C'eh-ta (1126-93), in his Kwei hai yii hen &", 4 mentions a. yen-Zi 
HI Bh tree, strong and fine, with a color like yen-Zi (that is, red), good 
for making arrowheads, and growing in Yuh eou, also in the caves of 
this department, and in the districts of Kwei-lin, in Kwan-si Province. 
A. Henry 5 gives for Yi-c'an in Se-£'wan a plant-name yen-Zi ma Ba 
M ("cosmetic hemp"), identified with Patrinia villosa. 

1 On account of the red color of the berries. 

2 See p. 478. 

1 Stuart, Chinese Materia Medica, p. 264; Matsumura, No. 2040; Perrot and 
Hurrier, Matiere m6dicale et pharmacopee sino-annamites, p. 116, where lo-k'wei 
is erroneously given as Chinese name of the plant. 

4 Ed. of Ci pu tsu cai ts'un Su, p. 28 b. 

s Chinese Names of Plants, p. 239 (Journal China Branch Roy. As. Soc, 
Vol. XXII, 1887). 


1 8. The Nan fan ts'ao mu huah W ~fS ^ fc %k, the oldest Chinese 
work devoted to the botany of southern China, attributed to Ki Han 
f& c^, a minister of the Emperor Hwei M (a.d. 290-309), contains 
the following notice: 1 — 

"The ye-si-mih $$ iS H flower and the mo-li ^ M flower (Jas- 
minum officinale, family Oleaceae) were brought over from western 
countries by Hu people tfi A, and have been planted in Kwah-tuh 
(Nan hai $J M) . The southerners are fond of their fragrant odor, and 
therefore cultivate them . . . The mo-li flower resembles the white 
variety of ts'iah-mi H J| (Cnidium monnieri), and its odor exceeds that 
of the ye-si-min" 

In another passage of the same work 2 it is stated that the U-kia 
Jb t flower (Lawsonia alba), 3 ye-si-min, and mo-li were introduced by 
Hu people from the country Ta Ts'in; that is, the Hellenistic Orient. 

The plant ye-si-min has been identified with Jasminum officinale; 
the plant mo-/*', with Jasminum sambac. Both species are now cultivated 
in China on account of the fragrancy of the flowers and the oil that 
they yield. 4 

The passage of the Nan fan ts'ao mu huah, first disclosed by Bret- 
schneider, 5 has given rise to various misunderstandings. Hirth 6 
remarked, "This foreign name, which is now common to all European 
languages, is said to be derived from Arabic-Persian jdsamin [read 
yasmin], and the occurrence of the word in a Chinese record written 
about a.d. 300 shows that it must have been in early use." Watters 7 
regarded yasmin as "one of the earliest Arabian words to be found in 
Chinese literature." It seems never to have occurred to these authors 

1 Ch. A, p. 2 (ed. of Han Wei ts'un Su). 

• Ch. B, p. 3. 

s See below, p. 334. 

4 The sambac is a favored flower of the Chinese. In Peking there are special 
gardeners who cultivate it exclusively. Every day in summer, the flower-buds are 
gathered before sunrise (without branches or leaves) and sold for the purpose of 
perfuming tea and snuff, and to adorn the head-dress of Chinese ladies. Jasminum 
officinale is not cultivated in Peking (Bretschneider, Chinese Recorder, Vol. Ill, 
1871, p. 225). 

6 Chinese Recorder, Vol. Ill, p. 225. 

6 China and the Roman Orient, p. 270. 

7 Essays on the Chinese Language, p. 354. 


330 Sino-Iranica 

that at this early date we know nothing about an Arabic or Persian 
language; and this rapprochement is wrong, even in view of the Chinese 
work itself, which distinctly says that both ye-si-min and mo-li were 
introduced from Ta Ts'in, the Hellenistic Orient. Pelliot 1 observes 
that the authenticity of the Chinese book has never been called into 
doubt, but expresses surprise at the fact that jasmine figures there 
under its Arabic name. But Arabic is surely excluded from the languages 
of Ta Ts'in. Moreover, thanks to the researches of L. Aurousseau, 2 
we now know that the Nan fan ts'ao mu cwan is impaired by inter- 
polations. The passage in question may therefore be a later addition, 
and, at all events, cannot be enlisted to prove that prior to the year 300 
there were people from western Asia in Canton. 3 Still less is it credible 
that, as asserted in the Chinese work, the Nan yiie kin ki *M & ff IE 
ascribed to Lu Kia ^ W, who lived in the third and second centuries 
B.C., should have alluded to the two species of Jasminum. 4 In fact, 
this author is made to say only that in the territory of Nan Yue the 
five cereals have no taste and the flowers have no odor, and merely 
that these flowers are particularly fragrant. Their names are not given, 
and it is Ki Han who refers them to ye-si-min and mo-li. It is out of 
the question that at the time of Lu Kia these two foreign plants should 
have been introduced over the maritime route into southern China; 
Lu Kia, if he has written this passage, may have as well had two other 
flowers in mind. 

The fact must not be overlooked, either, that the alleged introduction 
from Ta Ts'in is not contained in the historical texts relative to that 
country, nor is it confirmed by any other coeval or subsequent source. 

The Pei hu lu 5 mentions the flower under the names ye-si-mi ^1$ ^ 5Jf 
and white mo-li 6 3i ^'J ^6 as having been transplanted to China by 
Persians, like the pH-H-Sa or gold-coiri flower. 6 The Yu yan tsa tsu 
has furnished a brief description of the plant, 7 stating that its habitat 
is in Fu-lin and in Po-se (Persia). The Pen ts'ao kan mu, Kwah k'iin 
fan p l u, H and Hwa kin 9 state that the habitat of jasmine (mo-li) was 

1 Bull, de VEcolefranqaise, Vol. II, p. 146. 

2 See above, p. 263. 

3 Hirth, Chau Ju-kua, p. 6, note 1. 

* This point is discussed neither by Bretschneider nor by Hirth, who do not 
at all mention this reference. 

5 Ch. 3, p. 16 (see above, p. 268). 

6 See below, p. 335. 

7 Translated by Hirth, Journal Am. Or. Soc, Vol. XXX, 1910, p. 22. 

8 Ch. 22, p. 8 b. 
» Ch. 4, p. 9. 

Jasmine 331 

originally in Persia, and that it was thence transplanted into Kwan- 
tun. The first-named work adds that it is now (sixteenth century) 
cultivated in Yun-nan and Kwan-tun, but that it cannot stand cold, 
and is unsuited to the climate of China. The Tan k'ien tsun lu ^9" §& 
M $k of Yah Sen $§ W (1488-1559) is cited to the effect that "the name 
nai Iff used in the north of China is identical with what is termed in the 
Tsin Annals "H 1§F tsan nai hwa W ('hair-pin') §it ^B. 1 As regards this 
flower, it entered China a long time ago." 

Accordingly we meet in Chinese records the following names for 
jasmine : 2 — 

(1) ffl 5c* Hj ye-si-min, * ya-sit(si5)-mih, = Pahlavi yasmin, 
New Persian yasamin, yasmin, yasmun, Arabic yasmin, or I? ife 3li 
ye -si-mi, *ya-sit-mit (in Yu yan tsa tsu) = Middle Persian *yasmir (?). 3 
Judging from this philological evidence, the statement of the Yu yan 
tsa tsu, and Li Si-cen's opinion that the original habitat of the plant was 
in Persia, it seems preferable to think that it was really introduced from 
that country into China. The data of the Nan fan ts'ao mu twan are 
open to grave suspicion; but he who is ready to accept them is com- 
pelled to argue, that, on the one hand, the Persian term was extant in 
western Asia at least in the third century a.d., and that, on the other 
hand, the Indian word mallika (see No. 2) had reached Ta Ts'in about 
the same time. Either suggestion would be possible, but is not con- 
firmed by any West-Asiatic sources. 4 The evidence presented by the 
Chinese work is isolated; and its authority is not weighty enough, the 
relation of the modern text to the original issue of about a.d. 300 is 
too obscure, to derive from it such a far-reaching conclusion. The 
Persian- Arabic word has become the property of the entire world: all 
European languages have adopted it, and the Arabs diffused it along 
the east coast of Africa (Swahill yasmini, Madagasy dzasimini). 

(2) ~M f 'J or ^ ffi mo-li, b *mwat(mwal)-li=wa//t, transcription of 

1 This is the night-blooming jasmine (Nyctanthes arbor tristis), the musk-flower 
of India (Stuart, Chinese Materia Medica, p. 287). 

2 There are numerous varieties of Jasminum, — about 49 to 70 in India, about 
39 in the Archipelago, and about 15 in China and Japan. 

s From the Persian loan-word in Armenian, yasmik, Hubschmann (Armen. 
Gram., p. 198) justly infers a Pahlavi *yasmlk, beside yasmin. Thus also *yasmlt 
or *yasmlr may have existed in Pahlavi. 

4 It is noteworthy also that neither Dioscorides nor Galenus was acquainted 
with jasmine. 

5 For the expression of the element It are used various other characters which 
may be seen in the Kwan k'tin fan p'u (Ch. 22, p. 8 b); they are of no importance 
for the phonetic side of the case. 

332 Sino-Iranica 

Sanskrit mallikd (Jasminum sambac), Tibetan mal-li-ka, Siamese ma-li, 1 
Khmer maty or mlih, Cam molih. Malayan melati is derived from 
Sanskrit malati, which refers to Jasminum grandifiorum. Mongol 
melirge is independent. Hirth's identification with Syriac molo 2 must 
be rejected. 

(3) ffc $c san-mo, *san-mwat (Fukien mwak). This word is given 
in the Nan fan ts*ao mu cwah z as a synonyme of Lawsonia alba, furnish- 
ing the henna; but a confusion has here arisen, for the transcription 
does not answer to any foreign name of Lawsonia, but apparently cor- 
responds to Arabic zanbaq ("jasmine"), from which the botanical term 
sambac is derived. It is out of the question that this word was known 
to Ki Han: it is clearly an interpolation in his text. 

(4) M^ man hwa ("man flower") occurs in Buddhist literature, 
and is apparently an abridgment of Sanskrit sumana (Jasminum grandi- 
fiorum), which has been adopted into Persian as suman or saman. 

Jasminum officinale occurs in Kashmir, Kabul, Afghanistan, and 
Persia; in the latter country also in the wild state. 

Jasmine is discussed in Pahlavi literature (above, p. 192) and in the 
Persian pharmacopoeia of Abu Mansur. 4 C'an Te noticed the flower 
in the region of Samarkand. 6 It grows abundantly in the province of 
Fars in Persia. 9 

Oil of jasmine is a famous product among Arabs and Persians, being 
styled in Arabic duhn az-zanbaq. Its manufacture is briefly described in 
Ibn al-Baitar's compilation. 7 According to Istaxri, there is in the 
province of Darabejird in Persia an oil of jasmine that is to be found 
nowhere else. Sabur and Siraz were renowned for the same product. 8 

The oil of jasmine manufactured in the West is mentioned in the 
Yu yah tsa tsu as a tonic. It was imported into China during the Sung 
period, as we learn from the Wei lio WM-, 9 written by Kao Se-sun 
iti M M, who lived toward the end of thk, twelfth and in the beginning 
of the thirteenth century. Here it is stated, "The ye-si-mih flower is 
a flower of the western countries, snow-white in color. The Hu #J 
(Iranians or foreigners) bring it to Kiao-cou and Canton, and every one 

1 Pallegoix i Description du royaume Thai, Vol. I, p. 147. 
2 Journal Am. Or. Soc, Vol. XXX, 1910, p. 23. 

* Ch. B, p. 3. See below, p. 334. 
4 Achundow, Abu Mansur, p. 147. 

6 Bretschneider, Mediaeval Researches, Vol. I, p. 131. 

8 G. Le Strange, Description of the Province of Fars, p. 51. 

7 L. Leclerc, Trait6 des simples, Vol. II, p. ill. 

8 P. Schwarz, Iran, pp. 52, 94, 97, 165. 

• Ch. 9, p. 9. 

Jasmine 333 

is fond of its fragrance and plants this flower. According to the Kwan 
cou t'u kin £c '}W K l& ('Gazetteer of Kwan-tun Province'), oil of 
jasmine is imported on ships; for the Hu gather the flowers to press 
from them oil, which is beneficial for leprosy Jfit J&. 1 When this fatty 
substance is rubbed on the palm of the hand, the odor penetrates through 
the back of the hand." 

1 According to the Arabs, it is useful as a preventive of paralysis and epilepsy 
(Leclerc, /. c). 


19. It is well known that the leaves of Lawsonia alba or L. inermis, 
grown all over southern China, are extensively used by women and 
children as a finger-nail dye, and are therefore styled U kia hwa -JH ^P 
^6 ("finger-nail flower"). 1 This flower is mentioned in the Sanfu hwah 
t'u, 2 of unknown authorship and date, as having been transplanted 
from Nan Yue (South China) into the Fu-li Palace at the time of the 
Han Emperor Wu (140-87 B.C.)- This is doubtless an anachronism or 
a subsequent interpolation in the text of that book. The earliest datable 
reference to this plant is again contained in the Nan fan ts'ao mu hsoan by 
Ki Han, 3 by whom it is described as a tree from five to six feet in height, 
with tender and weak branches and leaves like those of the young elm- 
tree fiifc (Ulmus campestris), the flowers being snow-white like ye-si-mih 
and mo-li, but different in odor. As stated above (p. 329), this work goes 
on to say that these three plants were introduced by Hu people from 
Ta Ts'in, and cultivated in Kwah-tun. 4 The question arises again 
whether this passage was embodied in the original edition. It is some- 
what suspicious, chiefly for the reason that Ki Han adds the synonyme 
san-mo, which, as we have seen, in fact relates to jasmine. 

The Pei hu lu, h written about a.d. 875 by Twan Kun-lu, contains 
the following text under the heading H kia hwa: "The finger-nail flower 
is fine and white and of intense fragrance. The barbarians # A now 
plant it. Its name has not yet been explained. There are, further, the 
jasmine and the white mo-li. All these were transplanted to China by 
the Persians (Po-se). This is likewise the case with the pH-H-^a Rft/ 3 
^ (or 'gold coin') flower {Inula chinensis). Originally it was only 
produced abroad, but in the second year of the period Ta-t'un ^C IhI 
(a.d. 536 of the Liang dynasty) it came to China for the first time 
(#p 2fc vdb)." In the Yu yah tsa tsu, 6 written about fifteen years 
earlier, we read, "The gold-coin flower dk H 1fc, it is said, was originally 
produced abroad. In the second year of the period Ta-t'un of the 

1 Cf. Notes and Queries on China and Japan, Vol. I, 1867, pp. 40-41. Stuart, 
Chinese Materia Medica, p. 232. 

2 Ch. 3, p. 9 b (see above, p. 263). 

3 Ch. B, p. 3 (ed. of Han Wei ts'un I«). 

* Cf. also Hirth, China and the Roman Orient, p. 268. 
6 Ch. 3, p. 16 (see above, p. 268). 
6 Ch. 19, p. 10 b. 


Henna 335 

Liang (a.d. 536) it came to China. At the time of the Liang dynasty, 
people of Kin Sou M 'M used to gamble in their houses at backgammon 
with gold coins. When the supply of coins was exhausted, they resorted 
to gold-coin flowers. Hence Yu Huh $1 5A said, ' He who obtains flowers 
makes money.' " The same work likewise contains the following note: 1 
u PS~$i-SaWLP ^ is a synonyme for the gold-coin flower, 2 which was 
originally produced abroad, and came to China in the first year of 
the period Ta-t'uh of the Liang (a.d. 535)." The gold-coin flower vis- 
ualized by Twan Kuh-lu and Twan C'eh-§i assuredly cannot be Inula 
chinensis, which is a common, wild plant in northern China, and which 
is already mentioned in the Pie lu and by T'ao Huh-kih. 3 It is patent 
that this flower introduced under the Liang must have been a different 
species. The only method of solving the problem would be to determine 
the prototype of p'i-U-Sa, which is apparently the transcription of a 
foreign word. It is not stated to which language it belongs; but, judging 
from appearances, it is Sanskrit, and should be traceable to a form 
like *vislsa (or *vicesa). Such a Sanskrit plant-name is not to be 
found, however. Possibly the word is not Sanskrit. 4 

The Pei hu lu, accordingly, conceives the finger-nail flower as an 
introduction due to the Persians, but does not allude to its product, 
the henna. I fail to find any allusion to henna in other books of the 
T'ang period. I am under the impression that the use of this cosmetic 
did not come into existence in China before the Sung epoch, and that 
the practice was then introduced (or possibly only re-introduced) by 
Mohammedans, and was at first restricted to these. It is known that 
also the leaves of Impatiens balsamina {fun sien M, 'fill) mixed with alum 
are now used as a finger-nail dye, being therefore styled Ian U kia ts'ao 
tfe Jb ¥ W (" plant dyeing finger-nails"), 6 — a term first appearing 
in the Kiu hwan pen ts'ao, published early in the Ming period. The 
earliest source that mentions the practice is the Kwei sin tsa §i §£ ^r 

1 Ch. 19, p. 10 a. 

1 The addition of *f* before kin in the edition of Pai hai surely rests on an error. 

* Cf. also Bretschneider, Bot. Sin., pt. Ill, p. 158. 

4 The new Chinese Botanical Dictionary (p. 913) identifies the gold-coin 
flower with Inula britannica. In Buddhist lexicography it is identified with 
Sanskrit jdti (Jasminum grandiflorum; cf. Eitel, Handbook, p. 52). The same 
word means also "kind, class"; so does likewise vicesa, and the compound jati- 
vicesa denotes the specific characters of a plant (Hoernle, Bower Manuscript, 
p. 273). It is therefore possible that this term was taken by the Buddhists in 
the sense of "species of Jasminum," and that finally vicesa was retained as the 
name of the flower. 

6 Stuart, Chinese Materia Medica, p. 215; Pen ts'ao kan mu, Ch. 17 b, p. 12 b. 

336 Sino-Iranica 

H Ht 1 by Cou Mi ffl $? (1230-1320), who makes the following ob- 
servation: "As regards the red variety of the fun sien flower (Impatiens 
balsamina), the leaves are used, being pounded in a mortar and mixed 
with a little alum. 2 The finger-nails must first be thoroughly cleaned, 
and then this paste is applied to them. During the night a piece of 
silk is wrapped around them, and the dyeing takes effect. This process 
is repeated three or five times. The color resembles that of the yen-ci 
(Basella rubrum). Even by washing it does not come off, and keeps 
for fully ten days. At present many Mohammedan women are fond 
of using this cosmetic for dyeing their hands, and also apply it to cats 
and dogs for their amusement." The Pen ts'ao kan mu quotes only the 
last clause of this text. From what Cou Mi says, it does not appear 
that the custom was of ancient date; on the contrary, it does not seem 
to be older than the Sung period. 

None of the early Pen ts'ao makes mention of Lawsonia. It first 
appears in the Pen ts'ao kan mu. All that Li Si-cen is able to note 
amounts to this: that there are two varieties, a yellow and a white one, 
which bloom during the summer months; that its odor resembles that 
of must /fC JP (Osmanthus fragrans) ; and that it can be used for dyeing 
the finger-nails, being superior in this respect to the fun sien flower 
(Impatiens balsamina). Cen Kan-cufi JUS RlJ *H, an author of the Sung 
period, mentions the plant under the name i hian hwa ^1 & j& ("flower 
of peculiar fragrance"). 

It has generally been believed hitherto that the use of henna and 
the introduction of Lawsonia into China are of ancient date; but, in 
fact, the evidence is extremely weak. In my opinion, as far as the em- 
ployment of henna is concerned, we have to go down as far as the 
Sung period. It is noteworthy also that no foreign name of ancient date, 
either for the plant or its product, is on record. F. P. Smith and Stuart 
parade the term M |pj hai-na (Arabic hinna) without giving a reference. 
The very form of this transcription shows that it is of recent date: in 
fact, it occurs as late as the sixteenth century in the Pen ts'ao kan mu, z 
then in the K'iinfah p x u of 1630 4 and the Nun cen ts'iian $u H. Ifc d£r 1§F, 
published in 1619 by Su Kwah-k'i % it S*, the friend and supporter 
of the Jesuits. It also occurs in the Hwa kin of 1688. 5 

It is well known what extensive use of henna (Arabic hinna, hence 

1 Wk $k ±, p. 17 (ed. of Pat hai). 

2 In this manner the dye is also prepared at present. 

3 Ch. 17 b, p. 12 b. 

4 Kwa-h k'un fan p'u, Ch. 26, p. 4 b. The passages of the first edition are 
especially indicated. 

6 Ch. 5, p. 23 b. 

Henna 337 

Malayan inei) has been made in the west from ancient times. The 
Egyptians stained their hands red with the leaves of the plant 1 (Egyp- 
tian puqer, Coptic kuper or khuper, Hebrew kopher, Greek Kvirpos). All 
Mohammedan peoples have adopted this custom; and they even dye 
their hair with henna, also the manes, tails, and hoofs of horses. 2 The 
species of western Asia is identical with that of China, which is sponta- 
neous also in Baluchistan and in southern Persia. 3 Ancient Persia 
played a prominent r61e as mediator in the propagation of the plant. 4 
"They [the Persians] have also a custom of painting their hands, and, 
above all, their nails, with a red color, inclining to yellowish or orange, 
much near the color that our tanners nails are of. There are those 
who also paint their feet. This is so necessary an ornament in their 
married women, that this kind of paint is brought up, and distributed 
among those that are invited to their wedding dinners. They there- 
with paint also the bodies of such as dye maids, that when they appear 
before the Angels Examinants, they may be found more neat and 
handsome. This color is made of the herb, which they call Chinne, 
which hath leaves like those of liquorice, or rather those of myrtle. It 
grows in the Province of Erak, and it is dry'd, and beaten, small as 
flower, and there is put thereto a little of the juyce of sour pomegranate, 
or citron, or sometimes only fair water; and therewith they color their 
hands. And if they would have them to be of a darker color, they rub 
them afterwards with wall-nut leaves. This color will not be got off in 
fifteen days, though they wash their hands several times a day." 5 It 

1 V. Loret, Flore pharaonique, p. 80; Wcenig, Pflanzen im alten Aegypten, 
P- 349- 

* L. Leclerc, Traits des simples, Vol. I, p. 469; G. Jacob, Studien in arabischen 
Geographen, p. 172; A. v. Kremer, Culturgeschichte des Orients unter den Chalifen, 
Vol. II, p. 325. 

1 C. Joret, Plantes dans l'antiquite\ Vol. II, p. 47. 

4 Schweinfurth, Z. Ethnologie, Vol. XXIII, 1 891, p. 658. 

5 A. Olearius, Voyages of the Ambassadors to the Great Duke of Muscovy 
and the King of Persia (1633-39), P- 2 34 (London, 1669). I add the very exact 
description of the process given by Schlimmer (Terminologie, p. 343): "C'est avec 
la poudre fine des feuilles seches de cette plante, largement cultiv£e dans le midi 
de la Perse, que les indigenes se colorent les cheveux, la barbe et les ongles en rouge- 
orange. La poudre, form6e en pate avec de l'eau plus ou moins chaude, est appliquee 
sur les cheveux et les ongles et y reste pendant une ou deux heures, ayant soin de la 
tenir constamment humide en empfichant l'evaporation de son eau; apres quoi la 
partie est lavee soigneusement; l'effet de l'application du henna est de donner une 
couleur rouge-orange aux cheveux et aux ongles. Pour transformer cette couleur 
rougeatre en noir luisant, on enduit pendant deux ou trois autres heures les cheveux 
ou la barbe d'une seconde pate formee de feuilles pulv£ris6es finement d'une espece 
d'indigof ere, cultivee sur une large echelle dans la province de Kerman. Ces mani- 
pulations se pratiquent d'ordinaire au bain persan, ou la chaleur humide diminue 

338 Sino-Iranica 

seems more likely that the plant was transmitted to China from Persia 
than from western Asia, but the accounts of the Chinese in this case are 
too vague and deficient to enable us to reach a positive conclusion. 

In India, Lawsonia alba is said to be wild on the Coromandel coast. 
It is now cultivated throughout India. The use of henna as a cosmetic 
is universal among Mohammedan women, and to a greater or lesser 
extent among Hindu also; but that it dates "from very ancient times," 
as stated by Watt, 1 seems doubtful to me. There is no ancient Sanskrit 
term for the plant or the cosmetic (mendhi or mendhika is Neo-Sanskrit), 
and it would be more probable that its use is due to Mohammedan 
influence. Joret 2 holds that the tree, although it is perhaps indigenous, 
may have been planted only since the Mohammedan invasion. 3 

Francois Pyrard, who travelled from 1601 to 1610, reports the 
henna-furnishing plant on the Maldives, where it is styled innapa 
{—hlna-jai, "henna-leaf"). "The leaves are bruised," he remarks, 
"and rubbed on their hands and feet to make them red, which they 
esteem a great beauty. This color does not yield to any washing, nor 
until the nails grow, or a fresh skin comes over the flesh, and then (that 
is, at the end of five or six months) they rub them again." 4 

singulierement la dur6e de l'op6ration." While the Persians dye the whole of their 
hands as far as the wrist, also the soles of their feet, the Turks more commonly 
only tinge the nails; both use it for the hair. 

1 Commercial Products of India, p. 707. 

2 Plantes dans l'antiquitS, Vol. II, p. 273. 

8 Cf. also D. Hooper, Oil of Lawsonia alba, Journal As. Soc. Bengal, Vol. IV, 
1908, p. 35- 

* Voyage of F. Pyrard, ed. by A. Gray, Vol. II, p. 361 (Hakluyt Society). The 
first edition of this work appeared in Paris, 161 1. 


20. Under the term hu Vuh (Japanese koto) M fll ("t'ung tree of 
the Hu, Iranian Paulownia imperialist" that is, Populus balsamifera), 
the Annals of the Former Han Dynasty mention a wild-growing tree 
as characteristic of the flora of the Lob-nor region; for it is said to be 
plentiful in the kingdom of San-san HP H. 1 It is self-evident from the 
nomenclature that this was a species new to the Chinese, who discovered 
it in their advance through Turkistan in the second century B.C., but 
that the genus was somewhat familiar to them. The commentator 
Mon K'an states on this occasion that the hu Vun tree resembles the 
mulberry (Mortis alba), but has numerous crooked branches. A more 
elaborate annotation is furnished by Yen Si-ku (a.d. 570-645), who 
comments, "The hu Vun tree resembles the fun fll (Paulownia im- 
perialis), but not the mulberry; hence the name hu fun is bestowed 
upon it. This tree is punctured by insects, whereupon flows down a 
juice, that is commonly termed hu Vun lei ffl ffl M ('hu-t'un tears'), 
because it is said to resemble human tears. 2 When this substance 
penetrates earth or stone, it coagulates into a solid mass, somewhat on 
the order of rock salt, called wu-t'un kien 1o fil g (< natron of the wu-t'un 
tree,' Sterculia platanifolia). It serves for soldering metal, and is now 
used by all workmen." 3 

The T'un tien jil &, written by Tu Yu tt f6 between the years 
766 and 801, says that "the country Lou & 4 among the Si 2un ffi ^L 
produces an abundance of tamarisks Wzffl (Tamarix chinensis), hu t'un, 
and pai ts'ao Q ^ ('white herb or grass'), 5 the latter being eaten by 

1 Ts'ien Han Su, Ch. 96 A, p. 3 b. Cf. A. Wylie, Journal Anthropological In- 
stitute, Vol. X, 1881, p. 25. 

1 Pliny (xn, 18, § 33) speaks of a thorny shrub in Ariana on the borders of India, 
valuable for its tears, resembling the myrrh, but difficult of access on account of the 
adhering thorns (Contermina Indis gens Ariana appellatur, cui spina lacrima pretiosa 
murrae simili, difficili accessu propter aculeos adnexos). It is not known what plant 
is to be understood by the Plinian text; but the analogy of the "tears" with the 
above Chinese term is noteworthy. 

8 This text has been adopted by the T'ai p'in hwan yu ki (Ch. 181, p. 4) in 
describing the products of Lou-Ian. 

4 Abbreviated for Lou-Ian $£ jJQ, the original name of the kingdom of San-san. 

5 This is repeated from the Han Annals, which add also rushes. The "white 
grass " is explained by Yen Si-ku as " resembling the grass yu ^ (Setaria viridis), but 
finer and without awns; when dried, it assumes a white color, and serves as fodder 
for cattle and horses." 


340 Sino-Iranica 

cattle and horses. The hu fun looks as if it were corroded by insects. 
A resin flows down and comes out of this tree, which is popularly called 
'hu-t'un tears'. It can be used for soldering gold (or metal) and silver. 
In the colloquial language, they say also lii # instead of lei, which is 
faulty." 1 

The T'an pen ts'ao 1 is credited with this statement: "Hu fun lei 
is an important remedy for the teeth. At present this word is the name of 
a place west of Aksu. The tree is full of small holes. One can travel 
for several days and see nothing but hu fun trees in the forests. The 
leaves resemble those of the fun {Paulownia). The resin which is like 
glue flows out of the roots." 

The Lin piao lu i 3 states positively that hu fun lei is produced in 
Persia, being the sap of the hu fun tree, and adds that there are also 
"stone tears," H lei 15 M, which are collected from stones. 

Su Kuh, the reviser of the Pen ts'ao of the T'ang, makes this ob- 
servation: 4 "Hu fun lei is produced in the plains and marshes as well 
as in the mountains and valleys lying to the west of Su-cou HI #1. 
In its shape it resembles yellow vitriol {hwan fan 51 W), h but is far 
more solid. The worm-eaten trees are styled hu fun trees. When their 
sap filters into earth and stones, it forms a soil-made product like 
natron. This tree is high and large, its bark and leaves resembling those 
of the white poplar and the green fun ^ ffl. It belongs to the family 
of mulberries, and is hence called hu fun tree. Its wood is good for 
making implements." 

Han Pao-sen I? 'Pfc*^-, who edited the Su pen ts'ao 13 ^ ^ about 
the middle of the tenth century, states, "The tree occurs west of Lian- 
cou W. 'M (in Kan-su). In the beginning it resembles a willow; when 
it has grown, it resembles a mulberry and the fun. Its sap sinks into 
the soil, and is similar to earth and stone. It is used as a dye like the 
ginger-stone (kian Si ^ 3?) . 6 It is extremely salty and bitter. It is 
dissolved by the application of water, and then becomes like alum 
shale or saltpetre. It is collected during the winter months." 

Ta Mih ^C *$, who wrote a Pen ts'ao about a.d. 970, says with 
reference to this tree, "There are two kinds, — a tree-sap which is not 
employed in the pharmacopoeia, and a stone-sap collected on the 

1 Cf. Cen lei pen ts'ao, Ch. 13, p. 33. 

2 As quoted in the Ci wu min Si t'u k'ao, Ch. 35, p. 8 b. 

3 Ch. b, p. 7 a (see above, p. 268). 1 

4 Cen lei pen ts'ao, I.e. 

6 F. de Mely, Lapidaire chinois, p. 149. 

6 A variety of stalactite (see F. de Mely, Lapidaire chinois, p. 94; Geerts, 
Produits, p. 343; Cen lei pen ts'ao, Ch. 5, p. 32). 

The Balsam-Poplar 341 

surface of stones; this one only is utilize d as a medicine. It resembles 
in appearance small pieces of stone, and those colored like loess take 
the first place. The latter are employed as a remedy for toothache." 
Su Sun, in his T'u kin pen ts'ao, remarks that it then occurred among 
the Western Barbarians (Si Fan), and was traded by merchants. He 
adds that it was seldom used in the recipes of former times, but that 
it is now utilized for toothache and regarded as an important remedy in 

Li Si-cen 1 refers to the chapter on the Western Countries (Si yii 
£wan) in the Han Annals, stating that the tree was plentiful in the 
country Ku-si $ 8$ (Turf an). No such statement is made in the 
Annals of the Han with regard to this country, but, as we have seen, 
only with reference to San-san. 2 He then gives a brief resume' of the 
matter, setting down the two varieties of "tree-tears" and "stone- 

The Ming Geography mentions hu Vun lei as a product of Hami. 
The Kwan yii ki z notices it as a product of the Chikin Mongols between 
Su-cou and Sa-cou. The Si yii wen kien lu* written in 1777, states in 
regard to this tree that it is only good as fuel on account of its crooked 
growth: hence the natives of Turkistan merely call it odon or otun, 
which means "wood, fuel" in Turkish. 6 The tree itself is termed in 
Turk! tograk. 

The Hui k'ian Zi 6 likewise describes the hu fun tree of Hami, saying 
that the Mohammedans use its wood as fuel, but that some with 
ornamental designs is carved into cases for writing-brushes and into 

Bretschneider 7 has identified this tree with Populus euphratica, 
the wood of which is used as fuel in Turkistan. It is not known, however, 
that this tree produces a resin, such as is described by the Chinese. 
Moreover, this species is distributed through northern China; 8 while 
all Chinese records, both ancient and modern, speak of the hu t'un 

1 Pen ts'ao kan mu, Ch. 34, p. 22. 

2 There is a passage in the Swi kin lu where the hu t'un is mentioned, and may 
be referred to Ku-§i (Chavannes, T'oung Pao, 1905, p. 569). 

3 Above, p. 251. 

* Ch. 7, p. 9 (Wylie, Notes on Chinese Literature, p. 64). 

s This passage has already been translated correctly by W. Schott (Abh. Berl. 
Ak., 1842, p. 370). It was not quite comprehended by Bretschneider (Mediaeval 
Researches, Vol. II, p. 179), who writes, "The characters hu t'ung here are intended 
to render a foreign word which means 'fuel'." 

6 Above, p. 230. 

7 Mediaeval Researches, Vol. II, p. 179. 

8 Forbes and Hemsley, Journal Linnean Society, Vol. XXVI, p. 536. 

342 Sino-Iranica 

exclusively as a tree peculiar to Turkistan and Persia. The correct 
identification of the tree is Populus balsamifera, var. genuina Wesm. 1 
The easternmost boundary of this tree is presented by the hills of 
Kumbum east of the KukunCr, which geographically is part of Central 
Asia. The same species occurs also in Siberia and North America; it 
is called Hard by the French of Canada. It is met with, further, wild 
and cultivated, in the inner ranges of the north-western Himalaya, 
from Kunawar, altitude 8000 to 13000 feet, westwards. In western 
Tibet it is found up to 14000 feet. 2 The buds contain a balsam-resin 
which is considered antiscorbutic and diuretic, and was formerly im- 
ported into Europe under the name bawne facot and tacamahaca 3 com- 
munis (or vulgaris). Watt says that he can find no account of this 
exudation being utilized in India. It appears from the Chinese records 
that the tree must have been known to the Iranians of Central Asia 
and Persia, and we shall not fail in assuming that these were also the 
discoverers of the medical properties of the balsam. It is quite credible 
that it was efficacious in alleviating pain caused by carious teeth, as it 
would form an air-tight coating around them. 

1 Matsumura, Shokubutsu mei-i, No. 2518. 

* G. Watt, Dictionary of the Economic Products of India, Vol. VI, p. 325. 

1 The tacamahaca (a word of American-Indian origin) was first described 
by Nicoloso de Monardes (Dos libros el uno que trata de todas las cosas que traen 
de nuestras Indias Occidentales, Sevilla, 1569) : " Assi mismo traen de nueva Espafia 
otro genero de Goma, o resina, que llaman los Indios Tacamahaca. Y este mismo 
nombre dieron nuestros Espafioles. Es resina sacada por incision de un Arbol 
grande como Alamo, que es muy oloroso, echa el fruto Colorado como simiente de 
Peonia. Desta Resina o goma, usan mucho los Indios en sus enfermedades, mayor- 
mente en hinchazones, en qualquiera parte del cuerpo que se engendran, por que las 
ressuelue madura, y deshaze marauillosamente," etc. A copy of this very scarce work 
is in the Edward E. Ayer collection of the Newberry Library, Chicago; likewise 
the continuation Segunda parte del libro, He las cosas que se traen de nuestras 
Indias Occidentales (Sevilla, 1571). 


21. The word "manna," of Semitic origin (Hebrew man, Arabic 
mann), has been transmitted to us through the medium of Greek n&vva 
in the translation of the Septuaginta and the New Testament. Manna 
is a saccharine product discharged from the bark or leaves of a number 
of plants under certain conditions, either through the puncture of insects 
or by making incisions in the trunk and branches. Thus there are 
mannas of various nature and origin. The best-known manna is the 
exudation of Fraxinus ornus (or Ornus europaea), the so-called manna- 
ash, occurring in the Mediterranean region and Asia Minor. 1 The chief 
constituent of manna is manna-sugar or mannite, which occurs in 
many other plants besides Fraxinus. 

The Annals of the Sui Dynasty ascribe to the region of Kao-C'an 
i® H (Turfan) a plant, styled yah ts'e ^ M ("sheep-thorn"), the upper 
part of which produces honey of very excellent taste. 2 

C'en Ts'ah-k'i, who wrote in the first part of the eighth century, 
states that in the sand of Kiao-ho 3£ $f (Yarkhoto) there is a plant 
with hair on its top, and that in this hair honey is produced; it is styled 
by the Hu (Iranians) %n¥t. ( = fft) B. kHe-p % o-lo, *k'it(k'ir)-bwu5-la. 3 
The first element apparently corresponds to Persian xar ("thorn") or 
the dialectic form yar; 4 the second, to Persian burr a or bur a ("lamb"), 8 
so that the Chinese term yah ts'e presents itself as a literal rendering 
of the Persian (or rather a Middle-Persian or Sogdian) expression. 
In New Persian the term xar-i-lutur ("camel-thorn") is used, and, 
according to Aitchison, also xar-i-buzi ("goat's thorn"). 6 

It is noteworthy that the Chinese have preserved a Middle-Persian 
word for "manna," which has not yet been traced in an Iranian source. 
The plant (Hedysarum alhagi), widely diffused over all the arid lowlands 

1 Cf. the excellent investigation of D. Hanbury, Science Papers, pp. 355-368. 

5 Sui Su, Ch. 83, p. 3 b. The same text is also found in the Wei Su and Pei Si; 
in the T'ai p'in hwan yU ki (Ch. 180, p. 11 b) it is placed among the products of 
Ku-§i j|l iSft in Turfan. 

1 Stuart (Chinese Materia Medica, p. 258) erroneously writes the first char- 
acter ^q . He has not been able to identify the plant in question. 

4 P. Horn, Grundriss der iranischen Philologie, Vol. I, pt. 2, p. 70. 

5 In dialects of northern Persia also varre, varra, and werk (J. de Morgan, 
Mission en Perse, Vol. V, p. 208). 

6 Cf. D. Hooper, Journal As. Soc. Bengal, Vol. V, 1909, p. 33. 


344 Sino-Iranica 

of Persia, furnishes manna only in certain districts. Wherever it fails 
to yield this product, it serves as pasture to the camels (hence its name 
"thorn of camels"), and, according to the express assurance of Schlim- 
mer, 1 also to the sheep and goats. "Les indigenes des contrees de la 
Perse, ou se fait la r£colte de teren-djebin, me disent que les pasteurs 
sont obliges par les institutions communales de s'dloigner avec leurs 
troupeaux des plaines ou la plante mannifere abonde, parce que les 
moutons et chevres ne manqueraient de faire avorter la recolte." In 
regard to a related species {Hedysarum semenowi), S. KorZinski 2 
states that it is particularly relished by the sheep which fatten on it. 

The Lian se kuh tse ki He R9 & ■? Ifi 3 is cited in the Pen ts'ao kan mu 
as follows: "In Kao-S'ah there is manna (ts'e mi ffl m). Mr. Kie ifc 
5* says, In the town Nan-p'ih $3 2 F 4 ftft the plant yah ts'e is devoid of 
leaves, its honey is white in color and sweet of taste. The leaves of the 
plant yah ts'e in Salt City (Yen e'en Qft %,) are large, its honey is dark 
W in color, and its taste is indifferent. Kao-c'ah is the same as Kiao-ho, 
and is situated in the land of the Western Barbarians (Si Fan !§ #) ; 5 
at present it forms a large department (ta lou ^C 'M)." 

Wan Yen-te, who was sent on a mission to Turfan in a.d. 981, 
mentions the plant and its sweet manna in his narrative. 6 

Cou K'u-fei, who wrote the Lih wai tat ta in 11 78, describes the 
"genuine manna (sweet dew)" M ~W % of Mosul (#7 #T fcl Wu-se-li) 
as follows: 7 "This country has a number of famous mountains. When 
the autumn-dew falls, it hardens under the influence of the sun-rays 
into a substance of the appearance of sugar and hoar-frost, which is 
gathered and consumed. It has purifying, cooling, sweet, and nutritious 
qualities, and is known as genuine manna." 8 

Wan Ta-yuan & ~K $3, in his Too i U lio % 3^t J& £. of 1349, 9 has 

1 Terminologie, p. 357. 

2 Vegetation of Turkistan (in Russian), p. 77. 

3 The work of Can Yue (a.d. 667-730); see The Diamond, this volume, p. 6. 

4 Other texts write ^J 2, hu. 

8 This term, which in general denotes Tibet, but certainly cannot refer to Tibet 
in this connection, has evidently misled Stuart (Chinese Materia Medica, p. 258) 
into saying that the substance is spoken of as coming from Tangut. 

• Cf. W. Schott, Zur Uigurenfrage II, p. 47 (Abh. Berl. Akad., 1875). 

7 Ch. 3, p. 3 b (ed. of Ci pu tsu lai ts'wfi $u). Regarding the term kan lu, which 
also translates Sanskrit amrita, see Chavannes and Pelliot, Traits manicheen, 
P- 155. 

8 The same text with a few insignificant changes has been copied by Cao Zu-kwa 
(Hirth's translation, p. 140). 

9 Regarding this work, cf. Pelliot, Bull, de VEcole franqaise, Vol. IV, p. 255. 

Manna 345 

the following note regarding manna (kan lu) in Ma-k'o-se-li : * "Every 
year during the eighth and ninth months it rains manna, when the 
people make a pool to collect it. At sunrise it will condense like water- 
drops, and then it is dried. Its flavor is like that of crystallized sugar. 
They also store it in jars, mixing it with hot water, and this beverage 
serves as a remedy for malaria. There is an old saying that this is the 
country of the Amritaraja-tathagata ~W % rE iu 2&" 2 

Li Si-een, after quoting the texts of C'en Ts'an-k'i, the Pei Si, etc., 8 
arrives at the conclusion that these data refer to the same honey-bearing 
plant, but that it is unknown what plant is to be understood by the 
term yah ts'e. 

The Turk! name for this plant is yantaq, and the sweet resin accumu- 
lating on it is styled yantaq Sakari ("yantaq sugar"). 4 

The modern Persian name for the manna is tar-dngubin (Arabic 
terenjobin; hence Spanish tereniabin) ; and the plant which exudates the 
sweet substance, as stated, is styled xar-i-Sutur ("camel-thorn"). The 
manna suddenly appears toward the close of the summer during the 
night, and must be gathered during the early hours of the morning. It 
is eaten in its natural state, or is utilized for syrup (Sire) in Central Asia 
or in the sugar-factories of Meshed and Yezd in Persia. 5 The Persian 
word became known to the Chinese from Samarkand in the tran- 
scription ta-lah-ku-pin 3H 5P 1& St. 6 The product is described under 
the title kan lu ~H* % ("sweet dew") as being derived from a small 
plant, one to two feet high, growing densely, the leaves being fine like 
those of an Indigofera (Ian). The autumn dew hardens on the surface 
of the stems, and this product has a taste like sugar. It is gathered and 
boiled into sweetmeats. Under the same name, kan-lu, the Kwah yii ki 1 
describes a small plant of Samarkand, on the leaves of which accumu- 
lates in the autumn a dew as sweet in taste as honey, the leaves resem- 

1 Unidentified. It can hardly be identified with Mosul, as intimated by 

2 Rockhill, T*oung Pao, 191 5, p. 622. This Buddhist term has crept in here 
owing to the fact that kan lu ("sweet dew") serves as rendering of Sanskrit amrita 
("the nectar of the gods") and as designation for manna. 

5 Also the Yu yah tsa tsu, but this passage refers to India and to a different 
plant, and is therefore treated below in its proper setting. 

* A. v. Le Coq, Sprichw6rter und Lieder aus Turfan, p. 99. If the supposition 
of B. Munkacsi (Keleti szemle, Vol. XI, 1910, p. 353) be correct, that Hungarian 
gyanta (gydnta, jdnta, gyenta, "resin") and gyantdr ("varnish") may be Turkish 
loan-words, the above Turkl name would refer to the resinous character of the plant. 

5 VXmbery, Skizzen aus Mittelasien, p. 189. 

6 Ta Min i t'un U, Ch. 89, p. 23. 

7 Ch. 24, p. 26, of the edition printed in 1744; this passage is not contained in 
the original edition of 1600 (cf. above, p. 251, regarding the various editions). 

346 Sino-Iranica 

bling those of an Indigofera (Ian) ; and in the same work 1 this plant is 
referred to QarS-Khoja iK 'M under the name yan ts'e. Also the Ming 
Annals 2 contain the same reference. The plant in question has been 
identified by D. Hanbury with the camel-thorn (Alhagi camelorum) , 
a small spiny plant of the family Leguminosae, growing in Iran and 
Turkistan. 3 

In the fourteenth century, Odoric of Pordenone found near the 
city Huz in Persia manna of better quality and in greater abundance 
than in any part of the world. 4 The Persian-Arabic manna was made 
known in Europe during the sixteenth century by the traveller and 
naturalist Pierre Belon du Mons (i 5 18-64) > 5 who has this account: 
"Les Caloieres auoy6t de la Mane liquide recueillie en leurs montagnes, 
qu'ils appellent Tereniabin, a la difference de la dure: Car ce que les 
autheurs Arabes ont appelle" Tereniabin, est gard£e en pots de terre 
comme miel, et la portent vendre au Caire: qui est ce qu' Hippocrates 
nomma miel de Cedre, et les autres Grecs ont nomme* Ros6e du mont 
Liban: qui est differente a la Manne blanche seiche. Celle que nous 
auons en France, apport^e de Brianson, recueillie dessus les Meleses a 
la sommjte' des plus hautes montagnes, est dure, differente a la susdicte. 
Parquoy estant la Manne de deux sortes, Ion en trouve au Caire de 
l'vne et de l'autre es boutiques des marchands, expose*e en vente. 
L'vne est appellee Manne, et est dure: l'autre Tereniabin, et est liquide: 
et pource qu'en auons fait plus long discours au liure des arbres tousiours 
verds, n'en dirons autre chose en ce lieu." The Briancon manna men- 
tioned by Belon is collected from the larch-trees (Pinus larix) of south- 
ern France. 6 Garcia da Orta 7 described several kinds of manna, one 
brought to Ormuz from the country of the Uzbeg under the name 
xir quest or xircast, "which means the milk of a tree called quest, for xir 
[read Ur] is milk in the Persian language, so that it is the dew that falls 

1 Ch. 24, p. 6, of the original edition; and Ch. 24, p. 30 b, of the edition of 1744. 

3 Ch. 329 (cf. Bretschneider, Mediaeval Researches, Vol. II, p. 192). 

3 The plant is said to occur also in India (Sanskrit vigaladd and gandharl; that 
is, from Gandhara), Arabia, and Egypt, but, curiously, in those countries does not 
produce a sugar-like secretion. Consequently it cannot be claimed as the plant 
which furnished the manna to the Israelites in the desert (see the Dictionnaire de 
la Bible by F. Vigouroux, Vol. I, col. 367). The manna of northern India became 
known to the Chinese in recent times (see Lu fan kun Si 'k't fOi J|; ^ j£ $jC, p. 44, 
in TsHn lao fan ts'un Su). 

1 Yule, Cathay, new ed., Vol. II, p. 109; Cordier's edition of Odoric, p. 59. 

6 Les Observations de plusieurs singularitez, pp. 228-229 (Anvers, 1555). 

6 Fluckiger and Hanbury, Pharmacographia, p. 416. 

7 C. Markham, Colloquies, p. 280. 

Manna 347 

from these trees, or the gum that exudes from them. 1 The Portuguese 
corrupted the word to siracost." The other kind he calls tiriam-jabim 
or trumgibim (Persian tar-angubin). "They say that it is found among 
the thistles and in small pieces, somewhat of a red color. It is said that 
they are obtained by shaking the thistles with a stick, and that they are 
larger than a coriander-seed when dried, the color, as I said, between 
red and vermilion. The vulgar hold that it is a fruit, but I believe 
that it is a gum or resin. They think this is more wholesome than the 
kind we have, and it is much used in Persia and Ormuz." "Another 
kind comes in large pieces mixed with leaves. This is like that of Cala- 
bria, and is worth more money, coming by way of Bacora, a city of 
renown in Persia. Another kind is sometimes seen in Goa, liquid in 
leather bottles, which is like coagulated white honey. They sent this 
to me from Ormuz, for it corrupts quickly in our land, but the glass 
flasks preserve it. I do not know anything more about this medicine." 
John Fryer 2 speaks of the mellifluous dew a-nights turned into manna, 
which is white and granulated, and not inferior to the Calabrian. 
According to G. Watt, 8 shirkhist is the name for the white granular 
masses found in Persia on the shrub Cotoneaster nummularia; white 
taranjabin (= tar-angubin) is obtained from the camel-thorn (Alhagi 
camelorum and A. maurorum), growing in Persia, and consisting of a 
peculiar sugar called melezitose and cane-sugar. The former is chiefly 
brought from Herat, and is obtained also from Atraphaxis spinosa 
(Polygonaceae) . 4 

It is thus demonstrated also from a philological and historical point 
of view that the yah ts'e and k'ie-p'o-lo of the Chinese represent the 
species Alhagi camelorum. 

Another Persian name for manna is xoSkenjubtn, which means "dry 
honey." An Arabic tradition explains it as a dew that falls on trees in 
the mountains of Persia; while another Arabic author says, "It is dry 
honey brought from the mountains of Persia. It has a detestable odor. 
It is warm and dry, warmer and dryer than honey. Its properties in 
general are more energetic than those of honey." 5 This product, called 

1 Garcia's etymology is only partially correct. The Persian word is Hr-xeSt . 
which means "goat's milk." Hence Armenian SirixiSd, SirxeSd, HraxuSg, or hraxui 
(cf. E. Seidel, Mechithar, p. 210). 

1 New Account of East India and Persia, Vol. II, p. 201. 

* Agricultural Ledger, 1900, No. 17, p. 188. 

4 See FLtfcKiGER and Hanbury, op. cit., p. 415. According to Schlimmer 
(Terminologie, p. 357), this manna comes from Herat, Khorasan, and the district 
Lor- §ehrestanek. 

6 L. Leclerc, Traits des simples, Vol. II, p. 32. 

348 Sino-Iranica 

in India guzangabin, is collected from the tamarisk (Tamarix gallica, 
var. mannifera Ehrenb.) in the valleys of the Peninsula of Sinai and 
also in Persia. 1 In the latter country, the above name is likewise applied 
to a manna obtained from Astragalus florulentus and A. adscendens 
in the mountain-districts of Chahar-Mahal and Faraidan, and especially 
about the town of Khonsar, south-west of Ispahan. The best sorts of 
this manna, which are termed gaz-alefi or gaz-khonsar (from the prov- 
ince Khonsar), are obtained in August by shaking it from the branches, 
the little drops finally sticking together and forming a dirty, grayish- 
white, tough mass. According to Schlimmer, 2 the shrub on which this 
manna is formed is common everywhere, without yielding, however, 
the slightest trace of manna, which is solely obtained in the small 
province Khonsar or Khunsar. The cause for this phenomenon is 
sought in the existence there of the Coccus mannifer and in the absence 
of this insect in other parts of the country. Several Persian physicians 
of Ispahan, and some European authors, have attributed to the puncture 
of this insect the production of manna in Khonsar; and Schlimmer 
recommends transporting and acclimatizing the insect to those regions 
where Tamarix grows spontaneously. 

It has been stated that the earliest allusion to tamarisk-manna is 
to be found in Herodotus, 3 who says in regard to the men of the city 
Callatebus in Asia Minor that they make honey out of wheat and the 
fruit of the tamarisk. The case, however, is different; Herodotus does 
not allude to the exudation of the tree. 

Stuart 4 states that tamarisk-manna is called ten $u H£ %. The 
tamarisk belongs to the flora of China, three species of it being known. 5 
The Chinese, as far as I know, make no reference to a manna from any 
of these species; and the term pointed out by Stuart merely refers to 
the sap in the interior of the tree, which, according to the Pen ts'ao, is 
used in the Materia Medica. Ceh Tsiao, JIB W* of the Sung period, in 
bis T'un U jS iS, 6 simply defines l x eh %u as "the sap in the wood or 
trunk of the tamarisk." 7 

1 See particularly D. Hooper, Tamarisk Manna, Journal As. Soc. Bengal, 
Vol. V, 1909, pp. 31-36. 

2 Terminologie, p. 359. 
» vii, 31. 

4 Chinese Materia Medica, p. 259. 

8 Bretschneider, Bot. Sin., pt. II, No. 527; Pen ts'ao kan mu, Ch. 35 b, p. 9. 

6 Ch. 76, p. 12. 

7 The Turkl name for the tamarisk is yulgun. In Persian it is styled gaz or 
gazm (Kurd gazo or gezu), the fruit gazmazak or gazmazu (gaz basrah, the manna of 
the tree); further, balangmus't, balangmusk, or balanjmusk, and Arabic-Persian 

Manna 349 

There is, further, an oak-manna collected from Quercus vallonea 
Kotschy and Q. persica. These trees are visited in the month of August 
by immense numbers of a small white Coccus, from the puncture of 
which a saccharine fluid exudes, and solidifies in little grains. The people 
go out before sunrise, and shake the grains of manna from the branches 
on to linen cloths spread out beneath the trees. The exudation is also 
collected by dipping into vessels of hot water the small branches on 
which it is formed, and evaporating the saccharine solution to a syrupy 
consistence, which in this state is used for sweetening food, or is mixed 
with flour to form a sort of cake. 1 

Aside from the afore-mentioned mannas, Schlimmer 2 describes two 
other varieties which I have not found in any other author. One he 
calls in Persian Hker eighal ("sugar eighal"), saying that it is produced 
by the puncture of a worm in the plant. This worm he has himself 
found in fresh specimens. This manna is brought to Teheran by the 
farmers of the Elburs, Lawistan, and Dimawend, but the plant occurs 
also in the environment of Teheran and other places. Although this 
manna almost lacks sweetness, it is a remarkable pectoral and alleviates 
obstinate coughs. The other is the manna of Apocynum syriacutn, 
known in Persia as Hker al-ofr and imported from Yemen and Hedjaz. 
According to the Persian pharmacologists, it is the product of a 
nocturnal exudation solidified during the day, similar to small 
pieces of salt, either white, or gray, and even black. It is likewise 
employed medicinally. 

Manna belonged to the food-products of the ancient Iranians, and 
has figured in their kitchen from olden times. When the great king so- 
journed in Media, he received daily for his table a hundred baskets full 
of manna, each weighing ten mines. It was utilized like honey for 
the sweetening of beverages. 3 I am inclined to think that the Iranians 
diffused this practice over Central Asia. 

The Yu yah tsa tsu has a reference to manna of India, as follows: 
"In northern India there is a honey-plant growing in the form of a 
creeper with large leaves, without withering in the autumn and winter. 
While it receives hoar-frost and dew, it forms the honey." According 
to G. Watt, 4 some thirteen or fourteen plants in India are known to 

1 FlUckiger and Hanbury, Pharmacographia, p. 416; Hanbury, Science 
Papers, p. 287; Schlimmer (Terminologie, p. 358) attributes the oak-manna to the 
mountains of Kurdistan in Persia. 

1 Terminologie, p. 359. 

3 C. Joret, Plantes dans l'antiquite\ Vol. II, p. 93. Regarding manna in Persia, 
see also E. Seidel, Mechithar, p. 163. 

4 Commercial Products of India, p. 929. 

350 Sino-Iranica 

yield, under the parasitic influence of insects or otherwise, a sweet fluid 
called "manna." This is regularly collected and, like honey, enters more 
largely than sugar into the pharmaceutical preparations of the Hindu. 

The silicious concretion of crystalline form, found in the culms or 
joints of an Indian bamboo [Bambusa arundinacea) and known as 
tabashir, is styled in India also "bamboo manna," — decidedly a 
misnomer. On the other hand, a real manna has sometimes been 
discovered on the nodes of certain species of bamboo in India. 1 The 
subject of tabashir has nothing to do with manna, nor with Sino-Iranian 
relations; but, as the early history of this substance has not yet been 
correctly expounded, the following brief notes may not be unwelcome. 2 
Specimens of tabashir, procured by me in China in 1902, are in the 
American Museum of Natural History in New York. 3 

We now know that tabashir is due to an ancient discovery made in 
India, and that at an early date it was traded to China and Egypt. 
In recent years the very name has been traced in the form tabasis 
(rdjSao-is) in a Greek papyrus, where it is said that the porous stone is 
brought down [to Alexandria] from [upper] Egypt: the articles of 
Indian commerce were shipped across the Red Sea to the Egyptian 
ports, and then freighted on the Nile downward to the Delta. 4 The 
Indian origin of the article is evidenced, above all, by the fact that the 
Greek term tabasis (of the same phonetic appearance as Persian tabds'ir) 
is connected with Sanskrit tavak-kstrd (or tvak-ksira; ksira, "vegetable 
juice"), and permits us to reconstruct a Prakrit form tabas'ira; for the 
Greek importers or exporters naturally did not derive the word from 
Sanskrit, but from a vernacular idiom spoken somewhere on the west 
coast of India. Or, we have to assume that the Greeks received the 
word from the Persians, and the Persians from an Indian Prakrit. 5 

The Chinese, in like manner, at first imported the article from India, 
calling it "yellow of India" (T'ten-Zu hwan X ^ Jl). It is first men- 
tioned under this designation as a product of India in the Materia 
Medica published in the period K'ai-pao (a.d. 968-976), the K'ai pao 

1 See G. Watt, Agricultural Ledger, 1900, No. 17, pp. 185-189. 

2 The latest writer on the subject, G. F. Kunz (The Magic of Jewels and Charms, 
pp. 233-235, Philadelphia, 1915), has given only a few historical notes of mediaeval 

3 Cat. No. 70, 13834. This is incidentally mentioned here, as Dr. Kunz states 
that very little of the material has reached the United States. 

4 H. Diels, Antike Technik, p. 123. 

5 The Persian taba&r is first described by Abu Mansur (Achundow, p. 95), 
and is still eaten as a delicacy by Persian women (ibid., p. 247). In Armenian it is 

Tabashir 351 

pen ts'ao; but at the same time we are informed that it was then obtained 
from all bamboos of China, 1 and that the Chinese, according to their 
habit, adulterated the product with scorched bones, the arrowroot 
from Pachyrhizus angulatus, and other stuff. 2 The Pen ts'ao yen i of 
1116 s explains the substance as a natural production in bamboo, yellow 
like loess. The name was soon changed into "bamboo-yellow" {Zu 
hwan 1T jiaf) or "bamboo-grease" (fukao).* It is noticeable that the 
Chinese do not classify tabashir among stones, but conceive it as a 
production of bamboo, while the Hindu regard it as a kind of pearl. 

The earliest Arabic author who has described the substance is 
Abu Dulaf, who lived at the Court of the Samanides of Bokhara, and 
travelled in Central Asia about a.d. 940. He says that the product 
comes from MandQrapatan in northwestern India (Abulfeda and 
others state that Tana on the island of Salsette, twenty miles from 
Bombay, was the chief place of production) , and is exported from there 
into all countries of the world. It is produced by rushes, which, when they 
are dry and agitated by the wind, rub against one another; this motion 
develops heat and sets them afire. The blaze sometimes spreads over 
a surface of fifty parasangs, or even more. Tabashir is the product of 
these rushes. 8 Other Arabic authors cited by Ibn al-Baitar derive the 
substance from the Indian sugarcane, and let it come from all coasts 
of India; they dwell at length on its medicinal properties. 6 Garcia 
da Orta (1563), who was familiar with the drug, also mentions the 
burning of the canes, and states it as certain that the reason they set 
fire to them is to reach the heart; but sometimes they do not follow 
this practice, as appears from many specimens which are untouched 
by fire. He justly says that the Arabic name (tabaUr, in his Portuguese 
spelling tabaxir) is derived from the Persian, and means "milk or juice, 
or moisture." The ordinary price for the product in Persia and Arabia 
was its weight in silver. The canes, lofty and large like ash-trees, 

1 The Cen lei pen ts'ao (Ch. 13, p. 48) cites the same text from a work Lin hai 
& $8> '$£ t&< apparently an other work than the Lin hai i wu li mentioned by Bret- 
schneider (Bot. Sin., pt. 1, p. 169). 

3 The following assertion by Stuart (Chinese Materia Medica, p. 64)is erroneous: 
"The Chinese did not probably derive the substance originally from India, but it is 
possible that the knowledge of its medicinal uses were derived from that country, 
where it has been held in high esteem from very early times." The knowledge of 
this product and the product itself first reached the Chinese from India, and nat- 
urally induced them to search for it in their own bamboos. 

1 Ch. 14, p. 4 b (ed. of Lu Sin-yuan). 

4 Pen ts'ao kan mu, Ch. 37, p. 9. 

1 G. Ferrand, Textes relatifs a l'Extreme-Orient, p. 225. 
* L. Leclerc, Traits des simples, Vol. II, pp. 399-401. 

352 Sino-Iranica 

according to his statement, generate between the knots great humidity, 
like starch when it is much coagulated. The Indian carpenters, who 
work at these canes, find thick juice or pith, which they put on the lum- 
bar region or reins, and in case of a headache on the forehead; it is used 
by Indian physicians against over-heating, external or internal, and 
for fevers and dysentery. 1 The most interesting of all accounts remains 
that of Odoric of Pordenone (died in 1331), who, though he does not 
name the product and may partially confound it with bezoar, alludes 
to certain stones found in canes of Borneo, "which be such that if any 
man wear one of them upon his person he can never be hurt or wounded 
by iron in any shape, and so for the most part the men of that country 
do wear such stones upon them." 2 

J. A. de Mandelslo 3 gives the following notice of tabashir: "It 
is certain that on the coast of Malabar, Coromandel, Bisnagar, and 
near to Malacca, this sort of cane (called by the Javians mambu [bam- 
boo] ) produces a drug called sacar mambus, that is, sugar of mambu. 
The Arabians, the Persians, and the Moores call it tabaxir, which in 
their language signifies a white frozen liquor. These canes are as big 
as the body of a poplar, having straight branches, and leaves something 
longer than the olive-tree. They are divided into divers knots, wherein 
there is a certain white matter like starch, for which the Persians and 
Arabians give the weight in silver, for the use they make of it in physick, 
against burning feavers, and bloudy fluxes, but especially upon the first 
approaches of any disease." 

1 C. Markham, Colloquies of Garcia da Orta, pp. 409-414. A list of Sanskrit 
synonymes for tabashir is given by R. Schmidt (ZDMG, Vol. LXV, 1911, p. 745). 

2 Yule, Cathay, new ed. by Cordier, Vol. II, p. 161. 
8 Voyages and Travels, p. 120 (London, 1669). 



22. The riddles of asafoetida begin with the very name: there is no 
adequate explanation of our word asa or assa. The new Oxford English 
Dictionary ventures to derive it from Persian aza or aza. This word, 
however, means nothing but "mastic," a product entirely different 
from what we understand by asafoetida (p. 2 5 2) . In no Oriental language 
is there a word of the type asa or aza with reference to this product, so 
it could not have been handed on to Europe by an Oriental nation. 
Kaempfer, who in 1687 studied the plant in Laristan, and was fairly 
familiar with Persian, said that he was ignorant of the origin of the 
European name. 1 Littr£, the renowned author of the Dictionnaire 
francais, admits that the origin of asa is unknown, and wisely abstains 
from any theory. 2 The supposition has been advanced that asa was 
developed from the laser or laserpitium of Pliny (xix, 5), the latter 
having thus been mutilated by the druggists of the middle ages. 
This etymology, first given by Garcia da Orta, 8 has been indorsed 
by E. Borszczow, 4 a Polish botanist, to whom we owe an excellent 
investigation of the asa-furnishing plants. Although this explanation 
remains as yet unsatisfactory, as the alleged development from laser 
to asa is merely inferred, but cannot actually be proved from mediaeval 
documents, 6 it is better, at any rate, than the derivation from the 

Asafoetida is a vegetable product consisting of resin, gum, and 
essential oil in varying proportions, the resin generally amounting 
to more than one-half, derived from different umbelliferous plants, as 
Ferula narthex, alliacea, f&tida, persica, and scorodosma (or Scorodosma 

1 Amoenitates exoticae, p. 539. 

2 The suggestion has also been made that asa may be derived from Greek 
asi (?) ("disgust") or from Persian anguza (" asaf cetida ") ; thus at least it is said by 
F. Stuhlmann (Beitrage zur Kulturgeschichte Ostafrikas, p. 609). Neither is con- 
vincing. The former moves on the same high level as Li Si-cen's explanation of 
a-wei ("The barbarians call out a, expressing by this exclamation their horror at 
the abominable odor of this resin"). 

3 C. Markham, Colloquies, p. 41. John Parkinson (Theatrum botanicum, 
p. 1569, London, 1640) says, "There is none of the ancient Authours either Greeke, 
Latine, or Arabian, that hath made any mention of Asa, either dulcis or f cetida, 
but was first depraved by the Druggists and Apothecaries in forraigne parts, that in 
stead of Laser said Asa, from whence ever since the name of Asa hath continued." 

4 Mimoires de VAcad. de St. Pttersbourg, Vol. Ill, No. 8, i860, p. 4. 

5 DuCange does not even list the word "asafoetida." 


354 Sino-Iranica 

f&tidum). 1 It is generally used in India as a condiment, being espe- 
cially eaten with pulse and rice. Wherever the plant grows, the fresh 
leaves are cooked and eaten as a green vegetable, especially by the 
natives of Bukhara, who also consider as a delicacy the white under part 
of the stem when roasted and flavored with salt and butter. In the 
pharmacopoeia it is used as a stimulant and antispasmodic. 

Abu Mansur, the Persian Li Si-cen of the tenth century, discrimi- 
nates between two varieties of asafcetida (Persian anguyan, Arabic 
anjudan), a white and a black one, adding that there is a third kind 
called by the Romans sesalius. It renders food easily digestible, strength- 
ens the stomach, and alleviates pain of the joints in hands and feet. 
Rubbed into the skin, it dispels swellings, especially if the milky juice 
of the plant is employed. The root macerated in vinegar strengthens 
and purifies the stomach, promotes digestion, and acts as an appetizer. 2 

The Ferula and Scorodosma furnishing asafcetida are typically 
Iranian plants. According to Abu Hanlfa," asa grows in the sandy plains 
extending between Bost and the country Klkan in northern Persia. 
Abu Mansur designates the leaves of the variety from Sarachs near 
Merw as the best. Acgording to Istaxrl, asa was abundantly produced 
in the desert between the provinces Seistan and Makran; according to 
Edrlsi, in the environment of Kaleh Bust in Afghanistan. Kaempfer 
observed the harvest of the plant in Laristan in 1687, and gives the 
following notice on its occurrence: 4 "Patria eius sola est Persia, non 
Media, Libya, Syria aut Cyrenaica regio. In Persia plantam hodie 
alunt saltern duorum locorum tractus, videlicet campi montesque circa 
Heraat, emporium provinciae Chorasaan, et jugum montium in 
provincia Laar, quod a flumine Cuur adusque urbem Congo secundum 
Persici sinus tractum extenditur, duobus, alibi tribus pluribusve para- 
sangis a litore." Herat is a renowned place of production, presumably 
the exclusive centre of production at the present day, whence the 
product is shipped to India. 

The exact geographical distribution has been well outlined by E. 
Borszczow. 5 Aside from Persia proper, Scorodosma occurs also on the 
Oxus, on the Aral Sea, and in an isolated spot on the east coast of the 
Caspian Sea. Judging from Chinese accounts, plants yielding asa 
appear to have occurred also near Khotan (see below), Turf an, and 

1 The genus Ferula contains about sixty species. 

2 Achundow, Abu Mansur, p. 8. 

s Leclerc, Traite" des simples, Vol. I, p. 142. 

4 Amoenitates exoticae, p. 291. 

5 Ferulaceen der aralo-caspischen Wuste (Mimoires de VAcad. de St. PSters- 
bourg, Vol. Ill, No. 8, i860, p. 16). 


Shahrokia. 1 We do not know, however, what species here come into 

Cao 2u-kwa states that the home of asafoetida is in Mu-ku-lan 
/fc ^ SB, in the country of the Ta-§i (Ta-d2ik, Arabs). 2 Mu-ku-lan is 
identical with Mekran, the Gedrosia of the ancients, the Maka of 
the Old-Persian inscriptions. Alexander the Great crossed Gedrosia 
on his campaign to India, and we should expect that his scientific staff, 
which has left us so many valuable contributions to the flora of Iran 
and north-western India, might have also observed the plant furnishing 
asafoetida; in the floristic descriptions of the Alexander literature, how- 
ever, nothing can be found that could be interpreted as referring to 
this species. H. Bretzl 3 has made a forcible attempt to identify a 
plant briefly described by Theophrastus, 4 with Scorodosma joetidum; 
and A. Hort, 5 in his new edition and translation of Theophrastus, has 
followed him. The text runs thus: "There is another shrub [in Aria] 
as large as a cabbage, whose leaf is like that of the bay in size and 
shape. And if any animal should eat this, it is certain to die of it. 
Wherefore, wherever there were horses, they kept them under control" 
[that is, in Alexander's army]. This in no way fits the properties of 
Ferula or Scorodosma, which is non-poisonous, and does not hurt any 
animal. It is supposed also that the laserpitium or silphion and laser 
of Pliny 6 should, at least partially, relate to asafoetida; this, however, 
is rejected by some authors, and appears to me rather doubtful. Garcia 
da Orta 7 has already denied any connection between that plant of the 
ancients and asa. L. Leclerc 8 has discussed at length this much-dis- 
puted question. 

The first European author who made an exact report of asafoetida 

1 Bretschneider, Mediaeval Researches, Vol. II, pp. 193, 254. The inter- 
pretation of lu-wei ("rushes") as asafoetida in the Si yu ki (ibid., Vol. I, p. 85) seems 
to me a forced and erroneous interpretation. 

* Hirth and Rockhill, Chao Ju-kua, p. 224. 

3 Botanische Forschungen des Alexanderzuges, p. 285. 

4 Histor. plant., IV. iv, 12. 
s Vol. I, p. 321. 

6 xix, 15. The Medic juice, called silphion, and mentioned as a product of 
Media by Strabo (XL xm, 7), might possibly allude to a product of the nature 01 
asafoetida, especially as it is said in another passage (XV. 11, 10) that silphion grew 
in great abundance in the deserts of Bactriana, and promoted the digestion of the 
raw flesh on which Alexander's soldiers were forced to subsist there. According to 
others, the silphion of the ancients is Thapsia garganica (Engler, Pflanzenfamilien, 
Vol. Ill, pt. 8, p. 247). Regarding the Medic oil (oleum Medicum) see Ammianus 
Marcellinus, xxm, 6. 

7 C. Markham, Colloquies, p. 44. 

8 Traits des simples, Vol. I, p. 144. 

356 Sino-Iranica 

was Garcia da Orta in 1563. However, living and studying in Goa, 
India, he did not learn from what plant the product was derived. On 
its use in India he comments as follows: "The thing most used through- 
out India, and in all parts of it, is that Assa-fetida, as well for medicine 
as in cookery. A great quantity is used, for every Gentio who is able 
to get the means of buying it will buy it to flavor his food. The rich 
eat much of it, both Banyans and all the Gentios of Cambay, and he 
who imitates Pythagoras. These flavor the vegetables they eat with it; 
first rubbing the pan with it, and then using it as seasoning with every- 
thing they eat. All the other Gentios who can get it, eat it, and laborers 
who, having nothing more to eat than bread and onions, can only eat 
it when they feel a great need for it. The Moors all eat it, but in smaller 
quantity and only as a medicine. A Portuguese merchant highly praised 
the pot-herb used by these Banyans who bring this Assa-fetida, and 
I wished to try it and see whether it pleased my taste, but as I do not 
know our spinach very well, it did not seem so palatable to me as it 
did to the Portuguese who spoke to me about it. There is a respected 
and discreet man in these parts, holding an office under the king, who 
eats Assa-fetida to give him an appetite for his dinner, and finds it 
very good, taking it in doses of two drachms. He says there is a slightly 
bitter taste, but that this is appetising like eating olives. This is before 
swallowing, and afterwards it gives the person who takes it much con- 
tent. All the people in this country tell me that it is good to taste and 
to smell." 

Chr. Acosta or Da Costa 1 gives the following account: "Altiht, 
anjuden, Assa fetida, dulce y odorata medicina (de que entre los Doc- 
tores ha auido differentia y controuersia) es ona Goma, que del Coracone 
traen a Ormuz, y de Ormuz a la India, y del Guzarate y del reyno Dely 
(tierra muy fria) la qual por la otra parte confina con el Coracone, y con 
la region de Chiruan, como siente Auicena. Esta Goma es llamada de 
los Arabios Altiht, y Antit, y delos Indios Ingu, o Ingara. El arbol de 
adonde mana, se llama Anjuden, y otros le llaman Angeydan. 

" La Assa se aplica para leuatar el miembro viril, cosa muy vsada en 
aquellas partes : y no viene a proposito para la diminution del coito, vsar 
del tal cumo de Regaliza. Y en las diuisiones pone Razis Altiht por 
medicina para las fiestas de Venus: y Assa dulcis no la pone Doctor 
Arabe, ni Griego, ni Latino, que sea de autoridad, porque Regaliza 
se llama en Arabio Cuz, y el cumo del cozido, y reduzido en forma de 
Arrope, le llaman los Arabes Robalcuz, y los Espanoles corrompiendole 

1 Tractado de las drogas, y medicinas de las Indias orientales, p. 362 (Burgos, 


el nombre le llaman Rabacuz. De suerte que Robalcuz en Arabio, quiere 
dezir c.umo basto de Regaliza: porque Rob, es cumo basto, y Al, ar- 
ticulo de genitiuo, de, y Cuz, regaliza, y todo junto significa cumo 
basto de Regaliza: y assi no se puede Uamar a este cumo Assa dulcis. 
Los Indios la loan para el estomago, para facilitar el vientre, y para 
consumir las ventosidadas. Tambien curan con esta medicina los 
cauallos, que echan mucha ventosidad. En tanto tienen esta medicina 
que le llama aquella gente, principalmente la de Bisnaguer, manjar 
delos Dioses." 

John Fryer 1 relates, "In this country Assa Foetida is gathered at 
a place called Descoon; 2 some deliver it to be the juice of a cane or reed 
inspissated; others, of a tree wounded: It differs much from the stink- 
ing stuff called Hing, it being of the Province of Carmania: 3 This latter 
is that the Indians perfume themselves with, mixing it in all their pulse, 
and make it up in wafers to correct the windiness of their food, which 
they thunder up in belchings from the crudities created in their stom- 
achs; never thinking themselves at ease without this Theriac: And this 
is they cozen the Europeans with instead of Assa Faetida, of which 
it bears not only the smell, but color also, only it is more liquid." 

J. A. de Mandelslo 4 reports as follows: "The Hingh, which our 
dragsters and apothecaries call Assa fcetida, comes for the most part 
from Persia, but that which the Province of Utrad produces in the Indies 
is the best, and there is a great trafhck driven in it all over Indosthan. 
The plant which produces it is of two kinds; one grows like a bush, and 
hath small leaves, like rice, and the other resembles a turnip-leaf, and 
its greenness is like that of fig-tree leaves. It thrives best in stony and 
dry places, and its gum begins to come forth towards the latter end 
of summer, so that it must be gathered in autumn. The traffick of it 
is so much the greater in those parts, upon this account, that the 
Benjans of Guzuratta make use of it in all their sawces, and rub their 

1 New Account of East India and Persia, Vol. II, p. 195 (Hakluyt Soc, 1912). 

1 Kuh-i Dozgan, west of Kuristan. 

s Hing is mentioned by Fryer (Vol. I, p. 286) as in use among the natives of 
southern India, "to correct all distempers of the brain, as well as stomach," "a sort 
of liquid Assa Fcetida, whereby they smell odiously." This is the product of Ferula 
alliacea, collected near Yezd in Khorasan and in the province of Kerman, and 
chiefly used by the natives of Bombay (FlCckiger and Hanbury, Pharmacographia, 
pp. 319-320; Watt, Commercial Products of India, p. 534). Fryer's distinction be- 
tween hing and asafcetida shows well that there were different kinds and grades of 
the article, derived from different plants. Thus there is no reason to wonder that 
the Chinese Buddhist authors discriminate between hingu and a-wei (Chavannes 
and Pelliot, Traits manicheen, p. 234); the £'ou ts'ai ("stinking vegetable") is 
probably also a variety of this product. 

4 Voyages and Travels, p. 67 (London, 1669). 

358 Sino-Iranica 

pots and drinking vessels therewith, by which means they insensibly 
accustom themselves to that strong scent, which we in Europe are 
hardly able to endure." 

The Chinese understand by the term a-wei products of two different 
plants. Neither Bretschneider nor Stuart has noted this. Li Si-cen 1 
states that "there are two kinds of a-wei, — one an herb, the other a 
tree. The former is produced in Turkistan (Si yu), and can be sun- 
dried or boiled: this is the kind discussed by Su Kun. The latter is 
produced among the Southern Barbarians (Nan Fan), and it is the 
sap of the tree which is taken: this is the kind described by Li Sun, 
Su Sun, and C'en C'en." Su Kun of the T'ang period reports that 
11 a-wei grows among the Western Barbarians (Si Fan) and in K'un- 
lun. 2 Sprouts, leaves, root, and stems strongly resemble the pai li Q 
I3i (Angelica anomala). The root is pounded, and the sap extracted 
from it is dried in the sun and pressed into cakes. This is the first 
quality. Cut-up pieces of the root, properly dried, take the second 
rank. Its prominent characteristic is a rank odor, but it can also stop 
foul smells; indeed, it is a strange product. The Brahmans say that 
hUn-kii (Sanskrit hingu, see below) is the same as a-wei, and that the 
coagulated juice of the root is like glue; also that the root is sliced, 
dried in the sun, and malodorous. In the western countries (India) 
its consumption is forbidden. 3 Habitual enjoyment of it is said to do 
away with foul breath. The barbarians (3% A) prize it as the Chinese 
do pepper." This, indeed, relates to the plant or plants yielding asa, 
and Li Si-cen comments that its habitat is in Hwo oou (Qara-Khoja) 
and Sa-lu-hai-ya (Shahrokia). 4 Curiously enough, such a typical Iran- 
ian plant is passed over with silence in the ancient historical texts 
relative to Sasanian Persia. The only mention of it in the pre-T'ang 
Annals occurs in the Sui iw 5 with reference to the country Ts'ao iff 
north of the Ts'un-lin (identical with the Ki-pin of the Han), while 
the T x ai p'in hwan yil ki 6 ascribes a-wei to Ki-pin. 

The Yu yan tsa tsu 7 contains the following account of the product: 

1 Pen ts'ao kan mu, Ch. 34, p. 21. 

* K'un-lun is given as place of production in the Kwan li, written prior to 
a.d. 527, but there it is described as the product of a tree (see below). 

* It was prohibited to the monks of the Mahayana (cf . S. Levi, Journal asiatique, 
1915, I, p. 87). 

4 Bretschneider, Mediaeval Researches, Vol. II, pp. 253, 254, also 193. 

5 Ch. 83, p. 8 (also in the Pet Si). 
8 Ch. 182, p. 12 b. 

7 Ch. 18, p. 8 b. 


"A-wei is produced in Gazna #P 18 ffl> (*Gia-ja-na); 1 that is, in north- 
ern India. In Gazna its name is hin-yii (Sanskrit hingu). Its habitat 
is also in Persia, where it is termed a-yii-tsie (see below). The tree 
grows to a height of eight and nine feet. 2 The bark is green and yellow. 
In the third month the tree forms leaves which resemble a rodent's 
ear. It does not flower, nor does it produce fruit. The branches, when 
cut, have a continuous flow of sap like syrup, which consolidates, and 
is styled a-wei. The monk from the country Fu-lin, Wan ^ by name, 
and the monk from Magadha, T'i-p'o $i §1 (*De-bwa, Sanskrit Deva), 
agree in stating that the combination 3 of the sap with rice or beans, and 
powdered, forms what is called a-wei. " 4 

Another description of a-wei by the Buddhist monk Hwei Zi W. H , 
born in a.d. 680, has been made known by S. Lfevi. 6 The Chinese pil- 
grim points out that the plant is lacking in China, and is not to be seen 
in other kingdoms except in the region of Khotan. The root is as large 
as a turnip and white; it smells like garlic, and the people of Khotan 
feed on this root. The Buddhist pilgrim Yi Tsin, who travelled in 
a.d. 671-695, reports that a-wei is abundant in the western limit of 
India, and that all vegetables are mixed with it, clarified butter, oil, 
or any spice. 6 

Li Sun, who wrote in the second half of the eighth century, states 
that, "according to the Kwan li y a-wei grows in the country K'un-lun; 
it is a tree with a -sap of the appearance of the resin of the peach-tree. 
That which is black in color does not keep; that of yellow color is the 
best. Along the Yangtse in Yun-nan is found also a variety like the 
one imported in ships, juicy, and in taste identical with the yellow brand, 
but not yellow in color." Su Sun of the Sung period remarks that there 
is a-wei only in Kwan-Sou (Kwan-tun), and that it is the coagulated 
sap of a tree, which does not agree with the statement of Su Kuh. 
C'en C'en ffi. &, a distinguished physician, who wrote the Pen ts'ao 

1 In the Pen ts'ao kan mu, where the text is quoted from the Hai yao pen ts'ao 
of Li Sun, Persia is coupled with Gazna. Gazna is the capital of Jaguda, the Tsao- 
ku-C'a of Huan Tsan, the Zabulistan of the Arabs. Huan Tsan reported that 
asafoetida is abundant there (S. Julien, Memoires sur les contr^es occidentales. 
Vol. II, p. 187. Cf. S. Levi, Journal asiatique, 1915, I, p. 83). 

1 Thus in the text of the Pen ts'ao; in the edition of Pat hat: eighty or ninety 
feet. In fact, the sterns of Ferula reach an average height of from eight to ten feet. 

3 Instead of ifi\ of the text I read ^tl with the Pen ts'ao. 

* The translation of this passage by Hirth (Chau Ju-kua, p. 225) does not 
render the sense correctly. The two monks mean to say that the sap or resin is a 
condiment added to a dish of rice or beans, and that the whole mixture bears the 
name a-wei. 

5 Journal asiatique, 1915, 1, p. 89. 

6 Takakusu, I-tsing, pp. 128, 137. 

360 Sino-Iranica 

pie $wo about a.d. 1090, says, " A-wei is classed among trees. People 
of Kian-su and Ce-kian have now planted it. The odor of the branches 
and leaves is the same, but they are tasteless and yield no sap." The 
above K'un-lun refers to the K'un-lun of the Southern Sea; 1 and Li 
Si-cen comments that "this tree grows in Sumatra and Siam, and that 
it is not very high. The natives take a bamboo tube and stick it into 
the tree; the tube gradually becomes filled with the sap of the tree, and 
during the winter months they smash the tube and obtain the sap." 
Then he goes on to tell the curious tale of the sheep, in the same manner 
as Cao Zu-kwa. 2 

Cao Zu-kwa's notice that the resin is gathered and packed in skin 
bags is correct; for Garcia da Orta 3 reports that the gum, obtained 
by making cuts in the tree, is kept in bullock's hides, first anointed with 
blood, and then mixed with wheat flour. It is more difficult to account 
for the tradition given by the Chinese author, that, in order to neutralize 
the poison of the plant, a sheep is tied to the base of the tree and shot 
with arrows, whereupon the poison filters into the sheep that is doomed 
to death, and its carcass forms the asafcetida. This bit of folk-lore was 
certainly transmitted by Indian, Persian, or Arabic navigators, but any 
corresponding Western tradition has not yet been traced. Hobeich 
Ibn el-Hacen, quoted by Ibn al-Baitar, 4 insists on the poisonous action 
of the plant, and says that the harvests succeed in Sind only when asa 
is packed in a cloth and suspended at the mouth of water-courses, where 
the odor spread by the harvest will kill water-dogs and worms. Here 
we likewise meet the notion that the poisonous properties of the plant 
are capable of killing animals, and the sheep of the Chinese tradition 
is obviously suggested by the simile of white sheep-fat and the white 
vegetable fat of asa. In reality, sheep and goats are fond of the plant 
and fatten on it. 5 The asa ascribed to the country Ts'en-t'an in the Sun 
& 6 was surely an imported article. \ 

1 Not to the K'un-lun mountains, as assumed by Stuart (Chinese Materia 
Medica, p. 173). 

2 Needless to say, this Malayan asafoetida can have been but a substitute; but 
to what plant it refers, I am unable to say. The Tun si yah k'ao (Ch. 2, p. 18; 3, 
p. 6 b), published in 1618, mentions a-wei as product of Siam and Java. T'an Ts'ui 
U 2^, in his Tien hai yu hen U, written in 1799 (Ch. 3, p. 4, ed. of Wen yih lou yu 
ti ts'uh Su), states that the a-wei of Yun-nan is produced in Siam, being imported 
from Siam to Burma and brought from Burma up the Kin-sa kian. 

3 C. Markham, Colloquies, p. 47. 

4 Leclerc, Traits des simples, Vol. I, p. 447. 

5 E. Kaempfer, Amoenitates exoticae, p. 540; C. Joret, Plantes dans l'antiquite\ 
Vol. II, p. 100. 

6 Ch. 490; cf. Hirth, Chao Ju-kua, p. 127. I am not convinced that Ts'eh-t'an 
is identical with Ts'eh-pa or Zanguebar. 


In regard to the modern employment of the article, S. W. Williams 1 
writes, "It is brought from Bombay at the rate of $15 a picul, and 
ranks high in the Materia Medicaof the Chinese physician; it is exhibited 
in cholera, in syphilitic complaints and worms, and often forms an 
ingredient in the pills advertised to cure opium-smokers." It is chiefly 
believed, however, to assist in the digestion of meat and to correct the 
poison of stale meats (ptomaine poisoning), mushrooms, and herbs. 2 
In Annam it is carried in small bags as a preventive of cholera. 3 

The following ancient terms for asafoetida are on record: — 

(1) Persian P*J H£ ^St a-yii-tsie, *a-nu-zet= Middle Persian *anguzad; 
New Persian anguba, anguiad, anguydn, anguwan, anguddn, angiHak 
(stem angu-\-$ad = " gum" 4 ); Armenian ankutad, anjidan, Old Arme- 
nian anguZat, angtat; Arabic anjuddn. Garcia gives anjuden or angeidan 
as name of the tree from which asa is extracted. 

(2) Sanskrit HIS hin-kil, *hin-gu; % HI hin-yit, *hin-nu; H HI 
hiin-k'ii, *hun-gu; corresponding to Sanskrit hingu. In my opinion, 
the Sanskrit word is an ancient loan from Iranian. 6 Garcia gives imgo 
or imgara as Indian name, and forms with initial i appear in Indian 
vernaculars: cf. Telugu inguva; cf., further, Japanese ingu, Malayan 
angu (according to J. Bontitjs, who wrote in 1658, the Javanese and 
Malayans have also the word hin). 

(3) M 1% a-wei, *a-nwai; ;& S (in the Nirvana-sutra) yan-kwei, 
*an-kwai, correspond to an Indian or Iranian vernacular form of the 
type *ankwa or *ankwai, that we meet in Tokharian B or Kuca ankwa. 6 
This form is obviously based on Iranian angu, angwa. 

(4) Mongol "n "ai xa-si-ni (thus given as a Mongol term in the 
Pen ts'ao kan mu after the Yin San Zen yao of the Mongol period, written 
in 133 1), corresponds to Persian kasni, kisni, or gisnl ("asafoetida"), 
derived from the name of Gazni or Gazna, the capital of Zabulistan, 
which, according to Huan Tsah, was the habitat of the plant. A Mon- 
gol word of this type is not listed in the Mongol dictionaries of Kova- 
levski and Golstunski, but doubtless existed in the age of the Yuan, 

1 Chinese Commercial Guide, p. 80. 

2 Stuart, Chinese Materia Medica, p. 174. 

1 Perrot and Hurrier, Mat. m6d. et pharmacop^e sino-annamites, p. 161. 

4 Cf. Sanskrit jatuka (literally, "gum, lac ") = asafoetida. Hubschmann, Armen. 
Gram., p. 98. 

5 D'Herbelot (Bibliotheque orientale, Vol. I, p. 226; Vol. II, p. 327) derived 
the Persian word (written by him angiu, engiu, ingu; Arabic ingiu, ingudan) from 
Indian henk and hengu, ingu, for the reason that in India this drug is principally 
used; this certainly is not correct. 

N 6 Cf. T'oung Pao, 1915, pp. 274-275. 

362 Sino-Iranica 

when the Mongols introduced the condiment into China under that 
name, while they styled the root 18 M yin-lan. In modern Mongol, 
the name of the product is Ungun, which is borrowed from the Tibetan 
word mentioned below. 

In the Tibetan dialect of Ladakh, asafcetida is called kin or sip. 1 
The name sip or sup was reported by Falconer, who was the first to 
discover in 1838 Ferula narthex in western Tibet on the slopes of the 
mountains dividing Ladakh from Kashmir. 2 The word sip, however, 
is not generally Tibetan, but only of local value; in all probability, 
it is not of Tibetan origin. The common Tibetan word is Un-kun, 
which differs from the Iranian and Indian terms, and which, in view of 
the fact that the plant occurs in Tibetan regions, may be a purely Tibe- 
tan formation. 

Finally it may be mentioned that, according to Borszczow, 3 
Scorodosma is generally known to the inhabitants of the Aralo-Caspian 
territory . under the name sasyk-karai or keurok-kurai, which means 
as much as "malodorous rush." The Bukharans call it sasyk-kawar 
or simply kawar. 

1 Ramsay, Western Tibet, p. 7. 

2 Transactions Linnean Soc, Vol. XX, pt. 1, 1846, pp. 285-291. 

3 Op. cit., p. 25. 


23. There is only a single Chinese text relative to galbanum, which 
is contained in the Yu yah tsa tsu, 1 where it is said, "P'i-ts'i 1*$ J ^ 
(*bit-dzi, bir-zi, bir-zai) is a product of the country Po-se (Persia). 
In Fu-lin it is styled f 1 ill $J 4 han-p'o-li-Va (*xan-bwi5-li-da). 8 The 
tree grows to a height of more than ten feet, with a circumference of 
over a foot. Its bark is green, thin, and extremely bright. The leaves 
resemble those of the asafcetida plant (a-wei), three of them growing 
at the end of a branch. It does not flower or bear fruit. In the west- 
ern countries people are accustomed to cut the leaves in the eighth 
month; and they continue to do this more and more till the twelfth 
month. The new branches are thus very juicy and luxuriant; without 
the trimming process, they would infallibly fade away. In the seventh 
month the boughs are broken off, and there is a yellow sap of the 
appearance of honey and slightly fragrant, which is medicinally em- 
ployed in curing disease." 

Hirth has correctly identified the transcription pH-tsH with Persian 
birzai, which, however, like the other Po-se words in the Yu yah tsa tsu, 
must be regarded as Pahlavi or Middle Persian; 4 and the Fu-lin han- 
p'o-li-t'a he has equated with Aramaic xelbdnita, the latter from Hebrew 
xelbendh, one of the four ingredients of the sacred perfume (Exodus, 
xxx, 34-38). This is translated by the Septuaginta xo^fio.vr] and by 
the Vulgate galbanum. The substance is mentioned in three passages 

J Ch. 18, p. 11 b. 

2 Hirth, who is the first to have translated this text (Journal Am. Or. Soc. 
Vol. XXX, p. 21), writes this character with the phonetic element )§f, apparently 
in agreement with the edition of the Tsin tai pi Su; but this character is not author- 
ized by K'an-hi, and it is difficult to see how it could have the phonetic value p'i; 
we should expect ni. The above character is that given by K'an-hi, who cites under 
it the passage in question. It is thus written also in the Min hian p'u fe ^ §Jjf by 
Ye T'ih-kwei |j£ $3: £| (p. 10, ed. of Hian yen Is'un Su) and in the Pen ts'ao kan 
mu (Ch. 33, p. 6), where the pronunciation is explained by J5'J *biet. The editors 
of cyclopaedias were apparently staggered by this character, and most of them 
have chosen the phonetic man, which is obviously erroneous. None of our 
Chinese dictionaries lists the character. 

1 The Pen ts'ao kan mu (I. c.) annotates that the first character should have 
the sound ^ lo, *dwat, which is not very probable. 

4 There are also the forms ptrzed, barzed (Leclerc, Traite des simples, Vol. I, 
p. 201), berzed, barije, and bazrud; in India bireja, ganda-biroza. Another Persian 
term given by Schlimmer (Terminologie, p. 294) is weli. 


364 Sino-Iranica 

by Theophrastus: 1 it is produced in Syria from a plant called irkva^ 
("all-heal"); it is only the juice (6tt6s) which is called xo^fio-v-q, and 
which "was used in cases of miscarriage as well as for sprains and 
such-like troubles, also for the ears, and to strengthen the voice. The 
root was used in childbirth, and for flatulence in beasts of burden, 
further in making the iris-perfume (Iplvov fxvpov) because of its fra- 
grance; but the seed is stronger than the root. It grows in Syria, and 
is cut at the time of wheat-harvest." 2 

Pliny says that galbanum grows on the mountain Amanus in Syria 
as the exudation from a kind of ferula of the same name as the resin, 
sometimes known as stagonitis? Its medicinal employment is treated 
by him in detail. 4 Dioscorides 5 explains it as the gum of a plant which 
has the form of a ferula, growing in Syria, and called by some metopion. 
Abu Mansur 6 discusses the drug under the Arabic name quinna and the 
Persian name barzad. During the middle ages galbanum was well known 
in Europe from the fourteenth century onward. 7 

The philological result is confirmed by the botanical evidence, 
although Twan C'en-si's description, made from an oral report, not as 
an eye-witness, is naturally somewhat deficient; but it allows us to 
recognize the characteristics of a Ferula. It is perfectly correct that the 
leaves resemble those of the asafcetida Ferula, as a glance at the ex- 
cellent plates in the monograph of Borszczow (op. cit.) will convince 
one. It is likewise correct that the leaves grow at the ends of the twigs, 
and usually by threes. It is erroneous, however, that the tree does not 
flower or bear fruit. 8 The process of collecting the sap is briefly but 
well described. Nothing positive is known about the importation of gal- 
banum into China, although W. Ainslie 9 stated in 1826 that it was 

1 Histor. plant., IX. 1, 2; IX. vn, 2; IX* IX, 2. The term occurs also in the 
Greek papyri. V 

1 Cf. the new edition and translation of Theophrastus by A. Hort (Vol. II, 
p. 261). I do not see how the term "balsam of Mecca" (ibid., p. 219), which is a 
misnomer anyhow, can be employed in the translation of an ancient Greek 

1 Dat et galbanum Syria in eodem Amano monte e ferula, quae eiusdem nominis, 
resinae modo; stagonitim appellant (xn, 56, § 126). 

4 xxiv, 13. 

6 in, 87 (cf. Leclerc, Trait6 des simples, Vol. Ill, p. 115). 
8 Achundow, Abu Mansur, p. 108. 

7 See, for instance, K. v. Megenberg, Buch der Natur (written in 1349-50), 
ed. F. Pfeiffer, p. 367; Fluckiger and Hanbury, Pharmacographia, p. 321. 

8 The fruits are already mentioned by Theophrastus (Hist, plant., IX. ix, 2) 
as remedies. 

9 Materia Indica, Vol. I, p. 143. 

Galbanum 365 

sent from Bombay to China, and Stuart 1 regards this as entirely 
probable; but this is merely a supposition unsupported by any tangible 
data: no modern name is known under which the article might come. 
The three names given for galbanum in the English-Chinese Standard 
Dictionary are all wrong: the first, a-yil, refers to asafcetida (see above, 
p. 361) ; 2 the second, fm, denotes Liquidambar orientalis; and the third, 
pai sun hiah ("white pine aromatic"), relates to Pinus bungeana. 
The Pen ts'ao kah mu 3 has the notice on p'i-ts'i as an appendix to ' 'manna. " 
Li Si-cen, accordingly, did not know the nature of the product. He is 
content to cite the text of the Yu yah tsa tsu and to define the medical 
properties of the substance after C'en Ts'an-k'i of the T'ang. Only 
under the T'ang was galbanum known in China. 

The trees from which the product is obtained are usually identified 
with Ferula galbaniflua and F. rubricaulis or erubescens, both natives 
of Persia. The Syrian product used by the Hebrews and the ancients 
was apparently derived from a different though kindred species. 
F. rubricaulis, said by the botanist Buhse to be called in Persian khas- 
suih* is diffused all over northern Persia and in the Daena Mountains 
in the southern part of the country; it is frequent in the Demawend and 
on the slopes of the Alwend near Hamadan. 5 No incisions are made 
in the plant: the sap flowing out of the lower part of the stalks and from 
the base of the leaves is simply collected. The gum is amber-yellow, 
of not disagreeable, strongly aromatic odor, and soon softens between 
the fingers. Its taste is slightly bitter. Only in the vicinity of Hamadan, 
where the plant is exuberant, has the collecting of galbanum developed 
into an industry. 

Schlimmer 6 distinguishes two kinds, — a brown and a white-yel- 
lowish galbanum. The former (Persian bar zed or barije), the product of 
Ferula galbaniflua, is found near De Gerdon in the mountains Sa-ute- 
polagh between Teheran and Gezwin, in the valleys of Lars (Elburs), 
Khereghan, and Sawe, where the villagers gather it under the name 
balubu. The latter kind is the product of Dorema anchezi Boiss., en- 

1 Chinese Materia Medica, p. 181. 

* This is the name given for galbanum by F. P. Smith (Contributions towards 
the Materia Medica, p. ioo), but it is mere guesswork. 

» Ch. 33, p. 6. 

* Evidently identical with what Watt (Commercial Products of India, p. 535) 
writes khassnib, explaining it as a kind of galbanum from Shlraz. Loew (Aram. 
Pflanzennamen, p. 163) makes kassnih of this word. The word intended is apparently 
the kasni mentioned above (p. 361). 

5 Borszczow, op. cit., p. 35. 

* Terminologie, p. 295. 

366 Sino-Iranica 

countered by Buhse in the low mountains near Reshm (white galbanum) . 
Galbanum is also called kilyani in Persian. 

Borszczow has discovered in the Aralo-Caspian region another 
species of Ferula, named by him F. scha'ir from the native word lair 
(= Persian Sir, "milk- juice") for this plant. The juice of this species 
has the same properties as galbanum; also the plant has the same 

Abu Mansur 1 mentions a Ferula under the name sakbinaj (Arabic 
form, Persian sakbina), which his translator, the Persian physician 
Achundow, has identified with the Sagapenum resin of Ferula persica, 
said to be similar to galbanum and to be gathered in the mountains 
of Luristan. According to FLtJCKiGER and Hanbury, 2 the botanical 
origin of Sagapenum is unknown; but there is no doubt that this word 
{cxaykir-qvov in Dioscorides, in, 95, and Galenus; sacopenium in Pliny, 
xii, 56), in mediaeval pharmacy often written serapinum, is derived 
from the Persian word. 

The galbanum employed in India is imported from Persia to Bom- 
bay. Watt 3 distinguishes three kinds known in commerce, — Levant, 
Persian solid, and Persian liquid. The first comes from Shiraz, the 
second has an odor of turpentine, and the third is the gaoshir or jawa- 
shir; the latter being a yellow or greenish semi-fluid resin, generally 
mixed with the stems, flowers, and fruits of the plant. It is obtained from 
the stem, which, when injured, yields an orange-yellow gummy fluid. 
Generally, however, the galbanum of commerce forms round, agglu- 
tinated tears, about the size of peas, orange-brown outside, yellowish- 
white or bluish-green inside. The odor is not disagreeable, like that 
of asafcetida, and the taste is bitter. 

Galbanum consists of about 65 per cent resin, 20 per cent gum, and 
from 3 to 7 per cent volatile oil. I 

1 Achundow, Abu Mansur, p. 84. 

2 Pharmacographia, p. 342. 

3 Commercial Products of India, p. 535. 


24. Oak-galls (French noix de galles, Portuguese galhas) are globular 
excrescences caused by the gall-wasp {Cynips quercus jolii) puncturing 
the twigs, leaves, and buds, and depositing its ova in several species 
of oak (chiefly Quercus lusitanica var. injectoria), to be found in Asia 
Minor, Armenia, Syria, and Persia. In times of antiquity, galls were 
employed for technical and medicinal purposes. In consequence of 
their large percentage (up to 60 per cent) of tannic or Gallo-tannic 
acid, they served for tanning, still further for the dyeing of wool and 
the manufacture of ink. 1 Both Theophrastus 2 and Dioscorides 3 men- 
tion galls under the name /cr;/cis. Abu Mansur describes galls under 
the Arabic name afs* 

The greater part of the galls found in Indian bazars come from 
Persia, being brought by Arab merchants. 6 The Sanskrit name 
majuphala (phala, "fruit") is plainly a loan-word from the Persian 

In Chinese records, oak-galls are for the first time mentioned under 
the term wu-H-tse $& i£ 3F as products of Sasanian Persia. 6 They 
first became known in China under the T'ang from Persia, being intro- 
duced in the Materia Medica of the T'ang Dynasty (Paw pen ts'ao). 
The Paw pen lu m 3|£ $• states that they grow in sandy deserts, 7 and 
that the tree is like the tamarisk {Pen mm). A commentary, cited as 
kin cu ^ j£, adds that they are produced in Persia, while the Cen lei 
pen ts'ao* says that they grow in the country of the Western 2un 
(Iranians). The Yu yan tsa tsu 9 gives a description of the plant as 
follows: " Wu-H-tse M 15 :r are produced in the country Po-se (Persia), 

! BLtfMNER, Technologie, Vol. I, 2d ed., pp. 251, 268. 

1 Hist, plant., III. vm, 6. 

1 1, 146 (cf. Leclerc, Traits des simples, Vol. II, p. 457). See also Pliny, xm 
63; xvi, 26; xxiv, 109. 

4 Achundow, Abu Mansur, p. 98. 

6 W. Ainslie, Materia Indica, Vol. I, p. 145; Watt, Commercial Products of 
India, p. 911. 

• Sui Su, Ch. 83, p. 7 b. 

7 According to another reading, "in sandy deserts of the Western 2un" (that 
is,: Iranians). 

8 Ch. 14, p. 20. 

9 Ch. 18, p. 9. 


368 Sino-Iranica 

where they are styled Jf£ ft£ mo-tsei, ""mwa-dzak. 1 The tree grows to 
a height of from six to seven feet, 2 with a circumference of from eight to 
nine feet. The leaves resemble those of the peach, but are more oblong. 
It blossoms in the third month, the flowers being white, and their 
heart reddish. The seeds are round like pills, green in the beginning, 
but when ripe turning to yellow-white. Those punctured by insects 
and perforated are good for the preparation of leather; those without 
holes are used as medicine. This tree alternately produces galls one 
year and acorns (J$t M -?■ pa-lii tse, *bwa5-lu; Middle Persian *ballu, 
barru [see below], New Persian baluf), the size of a finger and three 
inches long, the next." 3 The latter notion is not a Chinese fancy, but 
the reproduction of a Persian belief. 4 

The Geography of the Ming (Ta Min i Vuh Zi) states that galls are 
produced in the country of the Arabs (Ta-si) and all barbarians, and 
that the tree is like the camphor-tree (Laurus camphor a), the fruits 
like the Chinese wild chestnuts (mao-li ^ W) . 

The Chinese transcriptions of the Iranian name do not "all repre- 
sent Persian mdzu," as reiterated by Hirth after Watters, but repro- 
duce older Middle-Persian forms. In fact, none of the Chinese render- 
ings can be the equivalent of mdzu. 

(i) J^ {$ (Yu yah tsa tsu) mo-tsei, *mwa-d2ak (dzak, zak), answers 
to a Middle Persian *mad2ak (madzak or mazak). 

(2) M ^ mo-Si, *mak-zak,= Middle Persian *maxzak. 

(3) M li wu-U, *mwu-zak, = Middle Persian *muzak. 

(4) $t ?J mu-H, *mut-zak, = Middle Persian *muzak. Compare 
with these various forms Tamil malakai, Telugu maUkai, and the 
magican of Barbosa. 

(5) J|l <$r 5 mo-t'u, *mwa-du,= Middle Persian *madu. 

^ $£ # to-mu-lii (in Cao 2u-kwa), *sa-mut-lwut, answers to Iranian 

1 Instead of tsei, some editions write %$ tso (*dzak, dzak), which is phonetically 
the same. 

2 The text has JJC, which should be corrected into /^, for the tree seldom rises 
higher than six feet. 

8 The text of the following last clause is corrupted, and varies in the different 
editions; it yields no acceptable sense. Hirth's translation (Chao Ju-kua, p. 215) 
is not intelligible to me. Watters (Essays on the Chinese Language, p. 349) is 
certainly wrong in saying that "the Chinese do not seem to know even yet the 
origin of these natural products" (oak-galls); this is plainly refuted by the above 
description. The T'u $u tsi I'en (XX, Ch. 310) and Ci wu min H t K u k'ao (Ch. 35, 
p. 21) even have a tolerably good sketch of the tree, showing galls on the leaves. 

4 E. Seidel, Mechithar, p. 127. 

6 The character 3fc £'a in Cao Zu-kwa, and thus adopted by Hirth (p. 215), is 
an error. 

Oak-Galls 369 

Zah-baluf ("the edible chestnut," Castanea vulgaris), which appears in 
the Bundahisn (above, p. 193), as correctly identified by Hirth; but 
fH M. p'u-lu and pa-lii of the Yu yan tsa tsu (see above) would indicate 
that the Chinese heard bulu and balu without a final *, and such forms 
may have existed in Middle-Persian dialects. In fact, we have this 
type in the dialect of the Kurd in the form berru, and in certain Kurd 
dialects baril and barru. 1 

1 Cf. J. de Morgan, Mission scientifique en Perse, Vol. V, p. 133. The Iranian 
term means literally "acorn of the Shah, royal acorn," somehow a certain analogy 
to Greek Aiis /SdXawrc ("acorn of Zeus"). The origin of Greek Kaar&vaiov or 
K&<rTavov is sought in Armenian kask ("chestnut") and kaskeni ("chestnut-tree"; 
see Schrader in Hehn, Kulturpflanzen, p. 402). According to the Armenian Geog. 
raphy of Moses of Khorene, the tree flourished in the Old-Armenian provinc e 
Duruperan (Daron); according to Galenus, near Sardes in Asia Minor; according to 
Daud, on Cyprus; according to Abu Mansur, also in Syria; while, according to the 
same author, Persia imported chestnuts from Adherbeijan and Arran; according to 
Schlimmer, from Russia (E. Seidel, Mechithar, p. 152). It is striking that the 
Chinese did not see the identity of the Iranian term with their /*' JJJi the common 
chestnut, several varieties of which grow in China. 


25. As indicated by our word "indigo" (from Latin indicum), this 
dye-stuff took its origin from India. The indigo-plant {Indigofera 
tinctoria), introduced into Persia from India, is discussed by Abu Man- 
sur under the name nil or Ilia. The leaves are said to strengthen the 
hair. The hair, if previously dyed with henna, becomes brilliant black 
from the pounded leaves of the plant. Another species, I. linifolia, 
is still used in Persia for dyeing beard and hair black. 1 The Persian 
words are derived from Sanskrit nila, as is likewise Arabic nilej. 2 Also 
nili hindi ("Indian indigo") occurs in Persian. Garcia da Orta has 
handed down a form anil* and in Spanish the plant is called anil 
(Portuguese and Italian anil). 4 It may be permissible to assume that 
indigo was first introduced into Sasanian Persia under the reign of 
Khosrau I AnOsarwan (a.d. 531-579); for Masudi, who wrote about 
a.d. 943, reports that this king received from India the book Kalila 
wa Dimna, the game of chess, and the black dye-stuff for the hair, 
called the Indian. 5 

Under the designation tsHn tai W H£ ("blue cosmetic for painting 
the eyebrows") the Chinese became acquainted with the true indigo 
and the Iranian practice mentioned above. The term is first on record 
as a product of Ts'ao iff (Jaguda) 6 and Ku-lan # B9 in the vicinity of 
Tokharestan; 7 during the T'ang period, the women of Fergana did not 
employ lead-powder, but daubed their eyebrows with tsHn tai. s Ma Ci 
of the tenth century says that "tsHn tai came from the country Po-se 
(Persia), but that now in T'ai-yuan, Lu-lih, Nan-k'ah, and other 

' Achundow, Abu Mansur, pp. 144, 271. Schlimmer (Terminologie, p. 395) 
gives ringi rl§ and wesme as Persian words for indigo-leaves. 

3 Leclerc, Traits des simples, Vol. Ill, p. 384. 

* C. Markham, Colloquies, p. 51. The form anil is also employed by F. Pyrard 
(Vol. II, p. 359, ed. of Hakluyt Society), who says that indigo is found only in the 
kingdom of Cambaye and Surat. 

4 Roediger and Pott (Z. /. Kunde d. Morg., Vol. VII, p. 125) regard this 
prefix o as the Semitic article (Arabic al-ml, an-nU). 

6 Barbier de Meynard and Pa vet de Courteille, Les Prairies d'or, Vol. II, 
p. 203. 

6 Sui Su, Ch. 83, p. 8 (see above, p. 317). 

7 T'ai p'in hwan yii ki, Ch. 186, p. 12. It was also found in Ki-pin {ibid., 
Ch. 182, p. 12 b). 

8 Ibid., Ch. 181, p. 13 b. 


Indigo 371 

places, a dye-stuff of similar virtues is made from tien j$ (the indigenous 
Polygonum tinctorium)." 1 Li Si-£en holds the opinion that the Persian 
tsHn tat was the foreign lan-tien Ml SS! (Indigo/era tinctoria). It must not 
be forgotten that the genus Indigojera comprises some three hundred 
species, and that it is therefore impossible to hope for exact identifica- 
tions in Oriental records. Says G. Watt 2 on this point, "Species of 
Indigojera are distributed throughout the tropical regions of the globe 
(both in the Old and New Worlds) with Africa as their headquarters. 
And in addition to the Indigoferas several widely different plants yield 
the self-same substance chemically. Hence, for many ages, the dye 
prepared from these has borne a synonymous name in most tongues, 
and to such an extent has this been the case that it is impossible to say 
for certain whether the nila of the classic authors of India denoted the 
self-same plant which yields the dye of that name in modern com- 
merce." "Indigo," therefore, is a generalized commercial label for a 
blue dye-stuff, but without botanical value. Thus also Chinese indigo 
is yielded by distinct plants in different parts of China. 3 

It is singular that the Chinese at one time imported indigo from 
Persia, where it was doubtless derived from India, and do not refer 
to India as the principal indigo-producing country. An interesting 
article on the term tsHn iai has been written by Hirth. 4 

1 Pen ts'ao kan mu, Ch. 16, p. 25 b. 

2 Commercial Products of India, p. 663. 

* Bretschneioer, Bot. Sin., pt. II, p. 212. 
4 Chinesische Studien, pp. 243-258. 


26. While rice is at present a common article of food of the Persian 
people, being particularly enjoyed as pilau, 1 it was entirely unknown 
in the days of Iranian antiquity. No word for "rice" appears in the 
Avesta. 2 Herodotus 3 mentions only wheat as the staple food of the 
Persians at the time of Cambyses. This negative evidence is signally 
confirmed by the Chinese annals, which positively state that there is 
no rice or millet in Sasanian Persia; 4 and on this point Chinese testi- 
mony carries weight, since the Chinese as a rice-eating nation were 
always anxious to ascertain whether rice was grown and consumed by 
foreign peoples. Indeed, the first question a travelling Chinese will 
ask on arrival at a new place will invariably refer to rice, its qualities 
and valuations. This is conspicuous in the memoirs of Can K'ien, 
the first Chinese who travelled extensively across Iranian territory, 
and carefully noted the cultivation of rice in Fergana (Ta-yuan), fur- 
ther for Parthia (An-si), and T'iao-ci (Chaldaea). The two last-named 
countries, however, he did not visit himself, but reported what he had 
heard about them. In the Sasanian epoch, Chinese records tell us 
that rice was plentiful in Kuca, Kasgar (Su-lek), Khotan, and Ts'ao 
(Jagucfe) north of the Ts'un-lin; 8 also in Si (Tashkend). 6 On the 
other hand, Aristobulus, a companion of Alexander on his expedition 
in Asia and author of an Alexander biography written after 285 B.C., 
states that rice grows in Bactriana, Babylonia, Susis, and in lower 
Syria; 7 and Diodorus 8 likewise emphasizes the abundance of rice in Susi- 

1 T'oung Pao, 19 16, p. 481. \ 

2 Modi, in Spiegel Memorial Volume, p. xxxvn. 

3 in, 22. 

* Wei $u, Ch. 102, pp. 5 b-6 a; Cou Su, Ch. 50, p. 6. Tabari (translation of 
Noldeke, p. 244) mentions rice among the crops taxed by Khusrau I (a.d. 531-578); 
but this is surely an interpolation, as in the following list of taxes rice is not men- 
tioned, while all other crops are. Another point to be considered is that in Arabic 
manuscripts, when the diacritical marks are omitted, the word birinj may be read 
as well naranj, which means "orange" (cf. Ouseley, Oriental Geography of Ebn 
Haukal, p. 221). 

6 Sui Su, Ch. 83, pp. 5 b, 7 b. 

* T'ai p'in hwan yii ki, Ch. 1 86, p. 7 b. 

* Strabo, XV. 1, 18. 
8 xix, 13. 


Rice 373 

ana. From these data Hehn 1 infers that under the rule of the Persians, 
and possibly inconsequence of their rule, rice-cultivation advanced from 
the Indus to the Euphrates, and that from there came also the Greek 
name 5pv£ a. This rice-cultivation, however, can have been but sporadic 
and along the outskirts of Iran; it did not affect Persia as a whole. The 
Chinese verdict of "no rice" in Sasanian Persia appears to me con- 
clusive, and it further seems to me that only from the Arabic period 
did the cultivation of rice become more general in Persia. This con- 
clusion is in harmony with the account of Hwi Cao sal M., a traveller 
in the beginning of the eighth century, who reports in regard to the 
people of Mohammedan Persia that they subsist only on pastry and 
meat, but have also rice, which is ground and made into cakes. 2 This 
conveys the impression that rice then was not a staple food, but merely 
a side-issue of minor importance. Yaqut mentions rice for the prov- 
inces Khuzistan and Sabur. 3 Abu Mansur, whose work is largely based 
on Arabic sources, is the first Persian author to discuss fully the subject 
of rice. 4 Solely a New-Persian word for "rice" is known, namely birinj 
or gurinj (Armenian and Ossetic brinj), which is usually regarded as a 
loan-word from Sanskrit vrlhi; Afghan vriie (with Greek 5puf o, (}pl£ a) 
is still nearer to the latter. In view of the historical situation, the 
reconstruction of an Avestan *verenja 5 or an Iranian *vrinji, 6 and the 
theory of an originally Aryan word for "rice," seem to me inadmissible. 

1 Kulturpflanzen, p. 505. 

* Hirth, Journal Am. Or. Soc, Vol. XXXIII, 1913, pp. 202, 204, 207. 

* B. de Meynard, Dictionnaire g£ographique de la Perse, pp. 217, 294. 

* Achundow, Abu Mansur, p. 5. J. Schiltberger (1396-1427), in his Bondage 
and Travels (p. 44, ed. of Hakluyt Society, 1879) speaks of the "rich country called 
Gilan, where rice and cotton alone is grown." 

6 P. Horn, Neupersische Etymologie, No. 208. 
6 H. HtJBSCHMANN, Persische Studien, p. 27. 


27. The pepper-plant (hu tsiao, Japanese ko$5, 1$ W>, Piper nigrum) 
deserves mention in this connection only inasmuch as it is listed among 
the products of Sasanian Persia. 1 Ibn Haukal says that pepper, sandal, 
and various kinds of drugs, were shipped from Siraf in Persia to all 
quarters of the world. 2 Pepper must have been introduced into Persia ' 
from India, which is the home of the shrub. 3 It is already enumerated 
among the plants of India in the Annals of the Han Dynasty. 4 The . 
Yu yan tsa tsu & refers it more specifically to Magadha, 6 pointing out 
its Sanskrit name marica or marica in the transcription &fc JH ^ mei- 
li-li? The term hu tsiao shows that not all plants whose names have 
the prefix hu are of Iranian origin: in this case hu distinctly alludes 
to India. 8 Tsiao is a general designation for spice-plants, principally 
belonging to the genus Zanthoxylon. Li Si-cen 9 observes that the black 
pepper received its name only for the reason that it is bitter of taste 
and resembles the tsiao, but that the pepper-fruit in fact is not a tsiao. 
It is interesting to note that the authors of the various Pen ts'ao seem 
to have lost sight of the fact of the Indian origin of the plant, and do 
not even refer to the Han Annals. Su Kuh states that hu tsiao grows 
among the Si Zun, which plainly shows that he took the word hu in 
the sense of peoples of Central Asia or Iranians, and substituted for it 

1 Suisu, Ch. 83, p. 7 b; Cou Su, Ch. 50, p. 6; and Weisu, Ch. 102, p. 6. According 
to Hirth (Chau Ju-kua, p. 223), this would mean that pepper was brought to China 
by Persian traders from India. I am unable to see this point. The texts in question 
simply give a list of products to be found in Persia, and say nothing about exporta- 
tion of any kind. V 

2 W. Ouseley, Oriental Geography of Ebn Haukal, p. 133. Regarding the for- 
mer importance of Siraf, which "in old times was a great city, very populous and 
full of merchandise, being the port of call for caravans and ships," see G. Le Strange, 
Description of the Province of Fars, pp. 41-43. 

8 In New Persian, pepper is called pilpil (Arabicized filfil, fulful), from the 
Sanskrit pippaM. 

* Hou Han Su, Ch. 118, p. 5 b. 

s Ch. 18, p. 11. 

8 Cf. Sanskrit magadha as an epithet of pepper. 

7 In fact, this form presupposes a vernacular type *meri&. 

8 Hu tsiao certainly does not mean "Western Barbarians (Tartar) pepper," 
as conceived by Waiters (Essays on the Chinese Language, p. 441). What had 
the "Tartars" to do with pepper? The Uigur adopted simply the Sanskrit word in 
the form murl. 

9 Pen ts'ao kan mu, Ch. 32, p. 3 b. 


Pepper 375 

its synonyme Si 2un; at least, it appears certain that the latter term 
bears no reference to India. Li Si-cen gives as localities where the 
plant is cultivated, "all countries of the Southern Barbarians (Nan 
Fan), Kiao-ci (Annam), Yun-nan, and Hai-nan." 

Another point of interest is that in the T'an pen ts'ao of Su Kun 
appears a species called San hu tsiao \\\ #) ffl> or wild pepper, described 
as resembling the cultivated species, of black color, with a grain the 
size of a black bean, acrid taste, great heat, and non-poisonous. This 
plant-name has been identified with Lindera glauca by A. Henry, 1 
who says that the fruit is eaten by the peasants of Yi-c'an, Se-6'wan. 
The same author offers a ye hu-tsiao ("wild pepper"), being Zanihoxy- 
lum setosum. 

Piper longum or Chavica roxburghii, Chinese ljh jc or iJfl pi-po, 
*pit-pat(pal), from Sanskrit pippali, is likewise attributed to Sasanian 
Persia. 2 This pepper must have been also imported into Iran from 
India, for it is a native of the hotter parts of India from Nepal east- 
ward to Assam, the Khasia hills and Bengal, westward to Bombay, 
and southward to Travancore, Ceylon, and Malacca. 3 It is therefore 
surprising to read in the Pen ts'ao of the T'ang that pi-po grows in the 
country Po-se: this cannot be Persia, but refers solely to the Malayan 
Po-se. For the rest, the Chinese were very well aware of the Indian 
origin of the plant, as particularly shown by the adoption of the San- 
skrit name. It is first mentioned in the Nan fan ts'ao mu twan, unless 
it be there one of the interpolations in which this work abounds, but 
it is mixed up with the betel-pepper (Chavica betel). 

1 Chinese Names of Plants, No. 45. 

2 Cou Su, Ch. 50, p. 6. 

3 Watt, Commercial Products of India, p. 891. 


28. The sugar-cane (Saccharum officinarum) is a typically Indian 
or rather Southeast-Asiatic, and merely a secondary Iranian culti- 
vation, but its history in Iran is of sufficient importance to devote here 
a few lines to this subject. The Sui Annals 1 attribute hard sugar 
(H-mi ^ 31*, literally, "stone honey") and pan-mi ^r 31* ("half honey") 
to Sasanian Persia and to Ts'ao (Jaguda). It is not known what kind 
of sugar is to be understood by the latter term. 2 Before the advent 
of sugar, honey was the universal ingredient for sweetening food-stuffs, 
and thus the ancients conceived the sugar of India as a kind of honey 
obtained from canes without the agency of bees. 3 The term U-mi first 
appears in the Nan fan ts'ao mu cwan, A which contains the first de- 
scription of the sugar-cane, and refers it to Kiao-cl (Tonking) ; according 
to this work, the natives of this country designate sugar as H-mi, which 
accordingly may be the literal rendering of a Kiao-£i term. In a.d. 285 
Fu-nan (Camboja) sent lu-1'6 ft M ("sugar-cane") as tribute to China. 5 
It seems that under the T'ang sugar was also imported from Persia 
China; for Moh Sen, who wrote the Si liao pen ts'ao in the second 
alf of the seventh century, says that the sugar coming from Po-se 
ersia) to Se-6'wan is excellent. Su Kun, the reviser of the T'an pen 
ts'ao of about a.d. 650, extols the sugar coming from the Si Zun, which 
may likewise allude to Iranian regions. Exact data as to the introduc- 
tion and dissemination of the sugar-cane in Persia are not available. 
E. O. v. Lippmann 6 has developed an elaborate theory to the effect that 

1 Sui Su, Ch. 83, p. 7 b. \ 

2 It is only contained in the Sui Su, not in the Wei Su (Ch. 102, p. 5 b), which 
has merely Si-mi. The sugar-cane was also grown in Su-le (Kashgar): T'ai p'ifi 
hwan yii ki, Ch. 181, p. 12 b. 

3 Pliny, xii, 17. 

4 Ch. 1, p. 4. 

6 This word apparently comes from a language spoken in Indo-China; it is already 
ascribed to the dictionary Swo wen. Subsequently it was replaced by kan "fj* 
("sweet") I'd or kan sp Id, presumably also the transcription of a foreign word. 
The Nan Ts'i Su mentions lu-1'6 as a product of Fu-nan (cf. Pelliot, Bull, de VEcole 
frangaise, Vol. Ill, p. 262). In C'i-t'u ffi J* (Siam) a wine of yellow color and fine 
aroma was prepared from sugar and mixed with the root of a Cucurbitacea (Sui Su, 
Ch. 82, p. 2 b). 

6 Geschichte des Zuckers, p. 93 (Leipzig, 1890); and Abhandlungen, Vol. I, 
p. 263. According to the same author, the Persians were the inventors of sugar- 
refining; but this is purely hypothetical. 


Sugar 377 

the Christians of the city Gundes&pur, which was in connection with 
India and cultivated Indian medicine, should have propagated the 
cane and promoted the sugar-industry. This is no more than an in- 
genious speculation, which, however, is not substantiated by any 
documents. The facts in the case are merely, that according to the 
Armenian historian Moses of Khorene, who wrote in the second half 
of the fifth century, sugar-cane was cultivated in Elymais near GundS- 
sapur, and that later Arabic writers, like Ibn Haukal, MuqaddasI, 
and Yaqut, mention the cultivation of the cane and the manufacture 
of sugar in certain parts of Persia. The above Chinese notice is of some 
importance in showing that sugar was known under the Sasanians in 
the sixth century. The Arabs, as is well known, took a profound inter- 
est in the sugar-industry after the conquest of Persia (a.d. 640), and 
disseminated the cane to Palestine, Syria, Egypt, etc. The Chinese 
owe nothing to the Persians as regards the technique of sugar-pro- 
duction. In a.d. 647 the Emperor T'ai Tsun was anxious to learn its 
secrets, and sent a mission to Magadha in India to study there the 
process of boiling sugar, and this method was adopted by the sugar- 
cane growers of Yan-eou. The color and taste of this product then were 
superior to that of India. 1 The art of refining sugar was taught the 
Chinese as late as the Mongol period by men from Cairo. 2 

1 T'an hut yao, Ch. ioo, p. 21. 

2 Yule, Marco Polo, Vol. II, pp. 226, 230. The latest writer on the subject of 
sugar in Persia is P. Schwarz (Der Islam, Vol. VI, 191 5, pp. 269-279), whose 
researches are restricted to the province of Ahwaz. In opposition to C. Ritter, who 
regarded Slraf on the Persian Gulf as the place whither the sugar-cane was first 
transplanted from India, he assigns this rdle to Hormuz; the first mention of refined 
sugar he finds in an Arabic poet of the seventh century. Lippmann's work is not 
known to him. 


29. The myrobalan Terminalia chebula, ho-li-lo M^W) (*ha-ri- 
lak, Japanese kariroku, Sanskrit haritaki, Tokharian arirak, Tibetan 
a-ru-ra, Newari halala; Persian halila, Arabic halila] and ihliligat) , was 
found in Persia. 1 The tree itself is indigenous to India, and the fruit 
was evidently imported from India into Persia. 2 This is confirmed by 
the fact that it is called in New Persian halila (Old Armenian halile), 
or halila-i kabuli, hinting at the provenience from Kabul. 3 

In the "Treatise on Wine," Tsiu p'uM H?, 4 written by Tou Kin M. W. 
of the Sung, it is said, "In the country Po-se there is a congee made 
from the three myrobalans (san-lo tsian =^Wi%fe), h resembling wine, and 
styled an-mo-lo $t M Wj (amalaka, Phyllanthus emblica) or p'i-li-lo 
Hit f& W} (vibhitaka, Terminalia belerica)." The source of this state- 
ment is not given. If Po-se in this case refers to Persia, it would go 
to show that the three myrobalans were known there. 

On the other hand, there is quite a different explanation of the 
tern, san-lo tsian. According to Ma Ci, who wrote in the tenth cen- 
tury, this i c the designation for a wine obtained from a flower of sweet 
flavor, growing in the countries of the West and gathered by the Hu. 
The name of the flower is 82 # t'o-te, *da-tik. 6 In this case the term 
san-lo may represent a transcription; it answers to ancient *sam-lak, 

1 Sui $u, Ch. 83, p. 7 b; Cou Su, Ch. 50, p. 6. 

2 Cf. T'oung Pao, 1915, pp. 275-276. Ho-li-lo were products of A-lo-yi-lo PSJ" 
Hi fp Hi in the north of Uddiyana (T'ai p%n hwan yii ki, Ch. 186, p. 12 b). 

3 Cf. G. Ferrand, Textes relatifs a l'Extr6me-Orient, p. 227. 

4 Ed. of T'an Sun ts'un $u, p. 20. 

5 The san lo are the three plants the names of which terminate in lo, — ho-li-lo 
(Terminalia chebula), p'i-li-lo (T. belerica, Sanskrit vibhitaka, Persian baUla), and 
a-mo-lo or an-mo-lo (Phyllanthus emblica, Sanskrit amalaka, Persian amola). 

6 The text is in the T'u Su tsi Ven, XX, Ch. 182, tsa hwa ts'ao pu, hui k'ao 2, 
p. 13 b. I cannot trace it in the Pen ts'ao kan mu. 



30. A fruit called yellow peach (hwan t'ao iH ffl&) or gold peach 
(kin t'ao 4£ #fc), of the size of a goose-egg, was introduced into China 
under the reign of the Emperor T'ai Tsun of the T'ang (a.d. 620-649), 
being presented by the country K'an Hfc (Sogdiana). 1 This introduction 
is assigned to the year 647 in the T'an hui yao, 2 where it is said that 
Sogdiana offered to the Court the yellow peach, being of the size of a 
goose-egg and golden in color, and hence styled also "gold peach." A 
somewhat earlier date for the introduction of this fruit is on record in 
the Ts'e fu yuan kwei* which has the notice that in a.d. 625 (under 
the Emperor Kao Tsu) Sogdiana presented gold peaches (kin t'ao) and 
silver peaches (yin t'ao), and that by imperial order they were planted 
in the gardens. This fruit is not mentioned in the Pen-ts'ao literature; 
it is not known what kind of fruit it was. Maybe it was a peculiar 
variety of peach. 


31. Fu-tse ffl -? is enumerated among the products of Sasanian 
Persia in the Sui £%. 4 Pai 6 fu-tse is attributed to the country Ts'ao 
(Jagu^a) north of the Ts'un-lih, 6 and to Eli-pin. 6 

In the form # -? fu-tse, it occurs in a prescription written on a 
wooden tablet of the Han period, found in Turkistan. 7 Fu-tse pft -J" is 
identified with Aconitum fischeri, cultivated on a large scale in Cah-min 
hien in the prefecture of Lu-han, Se-£'wan. 8 It is not known, however, 
that this species occurs in Persia. 

Yi Tsih calls attention to the fact that the medicinal herbs of India 
are not the same as those of China, and enumerates tubers of aconite 
together with fu-tse among the best drugs of China, and which are never 
found in India. 9 

1 Fun U wen kien ki, Ch. 7, p. I b (ed. of Ki fu ts'un Su). 

1 Ch. 200, p. 14; also T'ai p'in hwan yii ki, Ch. 183, p. 3. 

8 Ch. 970, p. 8 b. 

4 Ch. 83, p. 7 b; also Cou lu, Ch. 50, p. 6. 

6 Sui $u, ibid., p. 8 a. 

6 T'ai p'in hwan yii ki, Ch. 182, p. 12 b. 

7 Chavannes, Documents de l'epoque des Han, p. 115, No. 530. 

8 Stuart, Chinese Materia Medica, p. 10. 

9 Takakusu, Record of the Buddhist Religion, p. 148. 



32. Of the two species of mustard, Brassica or Sinapis juncea and 
S. alba, the former has always been a native of China {kiai ^r). The 
latter, however, was imported as late as the T'ang period. It is first 
mentioned by Su Kun in the Pen ts'ao of the T'ang (about a.d. 650) as 
coming from the Western Zuh (Si Zuh), 1 a term which, as noted, fre- 
quently refers to Iranian regions. In the Su pen ts'ao lij ^ ^, published 
about the middle of the tenth century by Han Pao-sen $$ ffi. Jr, we 
find the term ffll ^t* hu kiai ("mustard of the Hu"). C'en Ts'ah-k'i of 
the T'ang states that it grows in T'ai-yuan and Ho-tuh ?*f 3f€ (San-si), 
without referring to the foreign origin. Li Si-cen 2 annotates that this 
cultivation comes from the Hu and Zuh and abounds in Su (Se-6'wan), 
hence the names hu kiai and iw kiai ("mustard of Se-c'wan"), while 
the common designation is pai kiai ("white mustard"). This state 
of affairs plainly reveals the fact that the plant was conveyed to China 
over the land-route of Central Asia, while no allusion is made to an 
oversea transplantation. As shown by me on a previous occasion, 3 
the Si-hia word si-na ("mustard") appears to be related to 
Greek sinapi, and was probably carried into the Si-hia kingdom 
by Nestorian missionaries, who, we are informed by Marco 
Polo, were settled there. The same species was likewise foreign 
to the Tibetans, as is evidenced by their designation "white turnip" 
{yuns-kar). In India it is not indigenous, either: Watt 4 says that 
if met with at all, it occurs in gardens only within the tem- 
perate areas, or in upper India during the winter months; it is not 
a field crop. \ 

This genus comprises nearly a hundred species, all natives of the 
north temperate zones, and most of them of ancient European cultiva- 
tion (with an independent centre in China). 

Abu Mansur 5 distinguishes under the Arabic name karnab five kinds 
of Brassica, — Nabathaean, Brassica silvestris, B. marina, B. cypria 

1 The same definition is given by T'an Sen-wei in his Cen lei pen ts'ao (Ch. 27, 
P- 15). 

2 Pen ts'ao kan mu, Ch. 26, p. 12. 

3 T'oung Pao, 1915, p. 86. 

4 Commercial Products of India, p. 176. 
6 Achundow, Abu Mansur, p. no. 


Brassica 381 

(qanbit) and Syrian from Mosul. He further mentions Brassica rapa 
under the name Mgetn (Arabic Sal jam). 1 

33. One of the synonymes of yiin-Vai Wt ^ {Brassica rapa) is hu 
ts'ai #J 3£ ("vegetable of the Hu"). According to Li Si-cen, 2 this term 
was first applied to this vegetable by Fu K'ien ffii J^ of the second 
century his T'un su wen il f& 3$C. If this information were correct, 
this would be the earliest example of the occurrence of the term Hu in 
connection with a cultivated plant; but this Hu does not relate to 
Iranians, for Hu Hia ffll ?o , in his Pai pin fan If JfN j!j, a medical 
work of the Sui period (a.d. 589-618), styles the plant sai ts'ai ^^, 
which, according to Li Si-cen, has the same significance as hu ts'ai, and 
refers to IH #• Sai-wai, the Country beyond the Passes, Mongolia. 
Some even believe that Yun-t'ai is a place-name in Mongolia, where 
this plant thrives, and that it received therefrom its name. Such 
localities abstracted from plant-names are usually afterthoughts and 
fictitious. 3 The term yiin-Vai occurs in the early work Pie lu. 

Schlimmer 4 mentions Brassica capitata (Persian kalam pit), B. 
caulozapa (kalam gomri), and B. napus or rapa (Mgem). I have already 
pointed out that the Persians were active in disseminating species of 
Brassica and Raphanus to Tibet, the Turks, and Mongolia. 5 Reference 
has been made above (p. 199) to the fact that Brassica rapa (yiin-Vai) 
was introduced into China from Turkish tribes of Mongolia under the 
Later Han dynasty, and it would be reasonable to conclude that these 
had previously received the cultivation from Iranians. 6 Brassica rapa 
is very generally cultivated in Persia^ and most parts of India during 
the dry season, from October until March. 7 Yiin-Vai is enumerated 
among the choice vegetables of the country ^ fafc Mo-lu, *Mar-luk, in 
Arabia. 8 

The country of the Arabs produced the rape-turnip (man-tsin 
H H, Brassica rapa-depressa) with roots the size of a peck Pr t round, 
and of very sweet flavor. 9 

Yi Tsih, the Buddhist pilgrim of the seventh century, makes some 
commen>t on the difference between Indian and Chinese Brassica by saying, 

1 Achundow, Abu Mansur, p. 87. 

2 Pen ts'ao kan mu, Ch. 26, p. 9 b. 
* Compare p. 401. 

4 Terminologie, p. 93. 

5 T'oung Pao, 191 5, pp. 84, 87. 

8 The case would then be analogous to the history of the water-melon. 

7 W. Roxburgh, Flora Indica, p. 497. 

8 T'ai p'in hwan yu ki, Ch. 186, p. 16 b. 
» Ibid., Ch. 186, p. 15 b. 

382 Sino-Iranica 

"Man-tsin occurs [in India] in sufficient quantity and in two varieties, 
one with white, the other with black seeds. In Chinese translation it is 
called mustard (kie-tse iff* J~) . As in all countries, oil is pressed from it 
for culinary purposes. When eating it as a vegetable, I found it not 
very different from the man-tsin of China; but as regards the root, which 
is rather tough, ii is not identical with our man-tsin. The seeds are 
coarse, and again bear no relation to mustard-seeds. They are like those 
of Hovenia dulcis (U-kii ^R Wi) , transformed in their shape in conse- 
quence of the soil." 1 

1 This sentence is entirely misunderstood by J. Takakusu in his translation of 
Yi Tsih's work (p. 44), where we read, "The change in the growth of this plant is 
considered to be something like the change of an orange-tree into a bramble when 
brought north of the Yangtse River." The text has:^^i^R^@^3g^. 
There is nothing here about an orange or a bramble or the Yangtse. The character 
;f|j is erroneously used for Iff-, as is still the case in southern China (see Stuart, 
Chinese Materia Medica, p. 209), and ^ $| is a well-known botanical name for a 
rhamnaceous tree (not an orange), Hovenia dulcis. "Change of an orange-tree into 
a bramble" is nonsense in itself. 


34. Under the foreign term i^ H H-lo, *zi-la, the Chinese have 
not described the fennel (Foeniculum vulgare), as erroneously asserted 
by Watters 1 and Stuart, 2 but cummin {Cuminum cyminum) and 
caraway {Carum carui). This is fundamentally proved by the prototype, 
Middle Persian lira or zira, Sanskrit jira, of which H-lo (*zi-la) forms 
the regular transcription. 3 In India, jira refers to both cummin and 
caraway. 4 Although Cuminum is more or less cultivated in most prov- 
inces of India, except Bengal and Assam, there is, according to Watt, 
fairly conclusive evidence that it is nowhere indigenous; but in several 
I districts it would appear to be so far naturalized as to have been re- 
garded as "wild," even by competent observers. No doubt, it was 
transmitted to India from Iran. Cummin was known to the ancient 
Persians, being mentioned in the inscription of Cyrus at Persepolis, 6 
and at an early period penetrated from Iran to Egypt on the one hand, 
and to India on the other. 6 

Avicenna distinguishes four varieties of aimmin (Arabic kammiin), 7 
— that of Kirman, which is black; that of Persia, which is yellow and 
more active than the others; that of Syria, and the Nabathaean. 8 Each 
variety is both spontaneous and cultivated. Abu Mansur regards that 
of Kirman as the best, and styles it zire-i kirman? This name, accord- 
ing to Schlimmer, 10 would refer to caraway, also called zire-i siah, n 
while cummin is styled in Persian zire-i sebze or sefid. Caraway {Carum 

1 Essays on the Chinese Language, p. 440. He even adds "coriander," which 
is hu swi (p. 297). 

2 Chinese Materia Medica, p. 176. Fennel is hwi hian Jp] ^ff, while a synonyme 
of cummin is siao hwi hian ("small fennel"). 

s In the same form, the word occurs in Tibetan, zi-ra {T'oung Pao, 1916, p. 475). 
4 G. Watt, Commercial Products of India, p. 442. 

6 Joret, Plantes dans l'antiquit6, Vol. II, p. 66. 
8 Ibid., p. 258. 

7 Hebrew kammon, Assyrian katnanu, resulting in Greek kOuivov, Latin cumi- 
num, cyminum, or ciminum; Armenian caman; Persian kamun. 

8 Leclerc, Traits des simples, Vol. Ill, p. 196. 
• Achundow, Abu Mansur, pp. 112, 258. 

10 Terminologie, p. 112. 

11 In India, the Persian word siah refers to the black caraway (Carum bulbocasta- 
num), which confirms Schlimmer 's opinion. Also Avicenna's black cummin of 
Kirman apparently represents this species. This plant is a native of Baluchistan, 
Afghanistan, Kashmir, and Lahul, mainly occurring as a weed in cultivated land. 


384 Sino-Iranica 

carui), however, is commonly termed in Persian Sah-zire ("cummin of 
the Shah") or zlre-i ruml ("Byzantine or Turkish cummin"). 1 

While the philological evidence would speak in favor of a trans- 
mission of cummin from Persia to China, this point is not clearly brought 
out by our records. C'en Ts'an-k'i, who wrote in the first half of the 
eighth century, states that H-lo grows in Fu-si 1$ 1? (Bhoja, Sumatra). 
Li Sun, in his Hai yao pen ts'ao, says after the Kwan lou ki Ht m nS 
that the plant grows in the country Po-se; 2 and Su Sufi of the Sung 
notes that in his time it occurred in Lin-nan (Kwan-tuh) and adjoining 
regions. Now, the Kwan lou ki is said to have been written under the 
Tsin dynasty (a.d. 265-420) f and, as will be shown below in detail, the 
Po-se of Li Sun almost invariably denotes, not Persia, but the Malayan 
Po-se. Again, it is Li Sun who does not avail himself of the Iranian form 
s'i-lo=z'tra, but of the Sanskrit form jlraka, possibly conveyed through 
the medium of the Malayan Po-se. 

Li Si-cen has entered under U-lo another foreign word in the form 
3& H: Wti ts'e-mou-lo (*dzi-mu-lak), which he derived from the K'ai 
pao pen ts'ao, and which, in the same manner as U-lo, he stamps as a 
foreign word. This transcription has hitherto defied identification, 4 
because it is incorrectly recorded. It is met with correctly in the Cen 
lei pen ts'ao 5 in the form ^ W) ts'e-lo, *d2i-lak(rak), and this answers 
to Sanskrit jlraka. This form is handed down in the Hai yao pen ts'ao, 
written by Li Sun in the eighth century. Thus we have, on the one 
hand a Sanskrit form jlraka, conveyed by the Malayan Po-se to Kwan- 
tun in the T'ang period, and on the other hand the Iranian type H- 
lo=lira, which for phonetic reasons must likewise go back to the era 
of the T'ang, and which we should suppose had migrated overland to 
China. The latter point, for the time being, remains an hypothesis, 
which will perhaps be elucidated by the documents of Turkistan. 

1 Corresponding to Arabic karawya, the source of our word caraway. 

2 The Cen lei pen ts'ao (Ch. 13, p. 27 b) repeats this without citing a source. 
8 Cf . below, p. 475. 

4 Stuart, Chinese Materia Medica, p. 176. 
6 Ch. 13, p. 17 b. 


35. The Chinese records of the date-palm (Phoenix dactylifera) 
contain two points that are of interest to science: first, a contribution 
to the geographical distribution of the tree in ancient times; and, 
second, a temporary attempt at acclimating it in China. The tree is 
not indigenous there. It is for the first time in the T'ang period that 
we receive some information about it; but it is mentioned at an earlier 
date as a product of Sasanian Persia in both the Wei iw and Sui $u, 
under the name tsHen nien tsao ^ &£ 3t (" jujubes of thousand years," 
the jujube, Zizyphus vulgaris, being a native of China). 1 In the Yu yan 
tsa tsu, 2 the date is styled Pose tsao $t $f 31 ("Persian jujube"), with 
the observation that its habitat is in Po-se (Persia), or that it comes 
from there. 3 The Persian name is then given in the form 5H M k'u-man, 
*k'ut(k'ur)-man, which would correspond to a Middle Persian *xurman 
(*khurmang), Pazand and New Persian xurma, that was also adopted 
by Osmanli and Neo-Greek, xovpiias ("date") and Kovpnaorja. ("date- 
palm"), Albanian korme* The T'an £w 5 writes the same word $1 ^ 
hu-mah, *gu5(gur)-mah, answering to a Middle-Persian form *gurman 
or *kurmah. The New-Persian word is rendered ^ w M k'u-lu(ru)-ma 
in the Pen ts y ao kan mu; 6 this is the style of the Yuan transcriptions, 7 

1 This name was bestowed upon the tree, not, as erroneously asserted by[HiRTH 
(Chau Ju-kua, p. 210), "evidently on account of the stony hardness of the dates on 
reaching China," but, as stated in the Pen ts'ao kan mu (Ch. 31, p. 8), owing to the 
long-enduring character of the tree ^ $$ ^ ft" ^\ -{{£,. The same explanation 
holds good for the synonyme wan sui tsao ("jujube of ten thousand or numerous 
years "). Indeed, this palm lives to a great age, and trees of from one to two hundred 
years old continue to produce their annual crop. 

2 Ch. 18, p. 10. 

8 The same term, Po-se tsao, appears in a passage of the Pei hu lu (Ch. 2, p. 9 b), 
where the trunk and leaves of the sago-palm {Sago rumphii) are compared with those 
of the date. 

4 In Old Armenian of the fifth century we have the Iranian loan-word arntav, 
and hence it is inferred that the x of Persian was subsequently prefixed (HtJBSCH- 
mann, Persische Studien, p. 265; Armen. Gram., p. in). The date of the Chinese 
transcriptions proves that the initial * existed in Pahlavi. 

8 Ch. 221 b, p. 13. 

8 Ch. 31, p. 21. It is interesting to note that Li 5i-5en endeavors to make out 
a distinction between k'u-man and k'u-lu-ma by saying that the former denotes the 
tree, the latter the fruit; but both, in his opinion, are closely allied foreign words. 

7 The T'ang transcription, of course, is not "probably a distorted transcription 
of khurma," as asserted by Bretschneider (Chinese Recorder, 1871, p. 266), but, on 
the contrary, is very exact. 


386 Sino-Iranica 

and first occurs in the Co ken lu 15 $r" $$, published in 1366. The Persian 
word has also migrated into the modern Aryan languages of India, 
as well as into the Malayan group: Javanese kurma; Cam kuramo; 
Malayan, Dayak, and Sunda korma; Bugi and Makassar koromma; 
also into Khmer: rom'o, lom'd, amo. 

Following is the description of the tree given in the Yu yan tsa tsu: 
"It is thirty to forty feet in height, 1 and has a circumference of from 
five to six feet. The leaves resemble those of the fu fen dh W (a kind 
of rattan), and remain ever green. It blooms in the second month. 
The blossoms are shaped like those of the banana, and have a double 
bottom. They open gradually; and in the fissure are formed more than 
ten seed-cases, two inches long, yellow and white in color. When the 
kernel ripens, the seeds are black. In their appearance they resemble 
dried jujubes. They are good to eat and as sweet as candy." 

Another foreign word for the date is handed down by C'en Ts'ah-k'i 
in his Pen ts*ao U i, in the form M M wu-lou, *bu-nu. He identifies 
this term with the "Persian jujube," which he says grows in Persia, 
and has the appearance of a jujube. Li Si-6en annotates that the mean- 
ing of this word is not yet explained. Neither Bretschneider nor any 
one else has commented on this name. It is strikingly identical with 
the old Egyptian designation of the date, bunnu. 2 It is known that 
the Arabs have an infinite number of terms for the varieties of the date 
and the fruit in its various stages of growth, and it may be that they 
likewise adopted the Egyptian word and transmitted it to China. The 
common Arabic names are nakhl and tamr (Hebrew tamar, Syriac 
temar). On the other hand, the relation of wu-lou to the Egyptian word 
may be accidental, if we assume that wu-lou was originally the designa- 
tion of Cycas revoluta (see below), and was only subsequently trans- 
ferred to the date-palm. I 

The Lin piao lu i z by Liu Sun contains the following interesting 
account: — 

"In regard to the date ('Persian jujube'), this tree may be seen in 
the suburbs of Kwah-£ou (Canton). The trunk of the tree is entirely 
without branches, is straight, and rises to a height of from thirty to 
forty feet. The crown of the tree spreads in all directions, and forms 
over ten branches. The leaves are like those of the 'sea coir-palm' 

1 It even grows to a height of sixty or eighty feet. 

2 V. Loret, Flore pharaonique, p. 34. I concur with Loret in the opinion that 
the Egyptian word is the foundation of Greek #oiw£. The theory 'of Hehn (Kul- 
turpflanzen, p. 273) and upheld by Schrader {ibid., p. 284), that the latter might 
denote the Phoenician tree, does not seem to me correct. 

* Ch. b, p. 4 (see above, p. 268). 

The Date-Palm 387 

(hai tsun M tf , Chamaerops excelsa). 1 The trees planted in Kwah-cou 
bear fruit once in three or five years. The fruits resemble the green 
jujube growing in the north, but are smaller. They turn from green 
to yellow. When the leaves have come out, the fruit is formed in 
clusters, each cluster generally bearing from three to twenty berries, 
which require careful handling. The foreign as well as the domestic 
kind is consumed in our country. In color it resembles that of granulated 
sugar. Shell and meat are soft and bright. Baked into cakes or steamed 
in water, they are savory. The kernel is widely different from that of 
the jujube of the north. The two ends are not pointed [as in the jujube], 
but doubly rolled up and round like a small piece of red kino ^ $l. 2 
They must be carefully handled. When sown, no shoots sprout forth 
for a long time, so that one might suppose they would never mature." 

The date is clearly described in this text; and we learn from it that 
the tree was cultivated in Kwan-tuh, and its fruit was also imported 
during the T'ang period. As Liu Sun, author of that work, lived under 
the Emperor Cao Tsuii (a.d. 889-904), this notice refers to the end of 
the ninth century. 3 A. de Candolle 4 states erroneously that the 
Chinese received the tree from Persia in the third century of our era. 

In his note on the date, headed by the term wu-lou tse, Li Si-oen 5 
has produced a confusion of terms, and accordingly brought together 

1 In the text of this work, as cited in the Pen ts'ao kan mu, this clause is worded 
as follows: "The leaves are like those of the tsun-lii fS| fl?3 (Chamaerops excelsa), 
and hence the people of that locality style the tree [the date] hai tsun ('sea,' that is, 
'foreign coir-palm')." This would indeed appear more logical than the passage 
above, rendered after the edition of W u yin tien, which, however, must be regarded 
as more authoritative. Not only in this extract, but also in several others, does the 
Pen ts'ao kan mu exhibit many discrepancies from the Wu yin tien edition; this 
subject should merit closer study. In the present case there is only one other point 
worthy of special mention; and this is, that Li §i-cen, in his section of nomenclature, 
gives the synonyme ^§ He f an t sao ("foreign jujube") with reference to the Lin 
piao lu i. This term, however, does not occur in the text of this work as trans- 
mitted by him, or in the Wu yin tien edition. The latter has added a saying of the 
Emperor Wen j£ of the Wei dynasty, which has nothing to do with the date, and 
in which is found the phrase /L ^ fan tsao ("all jujubes"). In other editions, fan 
("foreign") was perhaps substituted for this fan, so that the existence of the 
synonyme established by Li and adopted by Bretschneider appears to be very 

2 See below, p. 478. 

* It is singular that Bretschneider, who has given a rather uncritical digest of 
the subject from the Pen ts'ao, does not at all mention this transplantation of the 
tree. To my mind, this is the most interesting point to be noted. Whether date- 
palms are still grown in Kwan-tun, I am not prepared to say; but, as foreign authors 
do not mention the fact, I almost doubt it. 

4 Origin of Cultivated Plants, p. 303. 

6 Pen ts'ao kan mu, Ch. 31, p. 8. 

388 Sino-Iranica 

a number of heterogeneous texts. Bretschneider 1 has accepted all this 
in good faith and without criticism. It is hardly necessary to be a 
botanist in order to see that the texts of the Nan fan ts'ao mu Zwan 
and Co ken lu, alleged to refer to the date, bear no relation to this tree. 2 
The hai tsao M HI described in the former work 3 may very well refer 
to Cycas revoluta* The text of the other book, which Bretschneider does 
not quote by its title, and erroneously characterizes as "a writer of the 
Ming," speaks of six "gold fruit" (kin kwo 4£ ^) trees growing in 
C'en-tu, capital of Se-6'wan, and, according to an oral tradition, planted 
at the time of the Han. Then follows a description of the tree, the 
foreign name of which is given as k'u-lu-ma (see above), and which, 
according to Bretschneider, suits the date-palm quite well. It is hardly 
credible, however, that this tree could ever thrive in the climate of 
Se-c'wan, and Bretschneider himself admits that the fruit of Salisburia 
adiantifolia now bears also the name kin kwo. Thus, despite the fact 
that the Persian name for the date is added, the passage of the Co ken 
lu is open to the suspicion of some misunderstanding. 

Not only did the Chinese know that the date is a product of Persia, 
but they knew also that it was utilized as food by certain tribes of the 

1 Chinese Recorder, 1871, pp. 265-267. 

2 Bretschneider, it should be understood, was personally acquainted with only 
the flora of Peking and its environment; for the rest, his familiarity with Chinese 
plants was mere book-knowledge, and botany as a science was almost foreign to 
him. Research in the history of cultivated plants was in its very beginning in 
his days; and his methods relating to such subjects were not very profound, and were 
rather crude. 

3 Ch. B, p. 4. Also Wu K'i-tsun, author of the Ci wu min H t'u k'ao (Ch. 17, 
p. 21), has identified the term wu-lou-tse with hai tsao. 

* Stuart, Chinese Materia Medica, p. 140; but Stuart falls into the other ex- 
treme by identifying with this species also, the terms Po-se tsao, ts'ien nien tsao, 
etc., which without any doubt relate to the^date. In Bretschneider's translation 
of the above text there is a curious misunderstanding. We read there, "In the year 
285 a.d. Lin-yi offered to the Emperor Wu-ti a hundred trees of the hai tsao. The 
prince Li-sha told the Emperor that in his travels by sea he saw fruits of this tree, 
which were, without exaggeration, as large as a melon." The text reads, "In the 
fifth year of the period T'ai-k'an (a.d. 284), Lin-yi presented to the Court a hundred 
trees. Li Sao-kun d|s /J> ;H* (the well-known magician) said to the Emperor Wu 
of the Han, ' During my sea- voyages I met Nan-k'i Sen t£; $8 ^ (the magician of 
the Blest Islands), who ate jujubes of the size of a gourd, which is by no means an 
exaggeration.' " The two events are not interrelated; the second refers to the second 
century B.C. Neither, however, has anything to do with the date. The working of 
Chinese logic is visibly manifest: the sea- travels of Li Sao-kun are combined with 
his fabulous jujube into the sea-jujube (hai tsao), and this imaginary product is 
associated with a real tree of that name. Li Si-£en's example shows at what fancies 
the Chinese finally arrive through their wrong associations of ideas; and Bret- 
schneider's example finally demonstrates that any Chinese data must first be taken 
under our microscope before being accepted by science. 

The Date-Palm 389 

East-African coast. The early texts relating to Ta Ts'in do not mention 
the palm; but at the end of the article Fu-lin (Syria), the T'an Su speaks 
of two countries, M ffl Mo-lin (*Mwa-lin, Mwa-rin) and ^ f# H 
Lao-p'o-sa (*Lav-bwi5-sar), as being situated 2000 li south-west of 
Fu-lin, and sheltering a dark-complexioned population. The land is 
barren, the people feed their horses on dried fish, and they themselves 
subsist on dates. 1 Bretschneider 2 was quite right in seeking this 
locality in Africa, but it is impossible to accept his suggestion that 
"perhaps the Chinese names Mo-lin and Lao-p'o-sa are intended to 
express the country of the Moors (Mauritania) or Lybia." Hirth 3 
did not discuss this weak theory, and, while locating the countries 
in question along the west coast of the Red Sea, did not attempt to 
identify the transcriptions. According to Ma Twan-lin, the country 
Mo-lin is situated south-west of the country ?& H H Yan-sa-lo, which 
Hirth tentatively equated with Jerusalem. This is out of the question, 
as Yah-sa-lo answers to an ancient Ah-sa5(sar)-la(ra). 4 Moreover, it 
is on record in the T % ai pHn hwan yii ki 5 that Mo-lin is south-west of 
# II $1 P'o-sa-lo (*Bwi5-sa5-la), so that this name is clearly identical 
with that of Ma Twan-lin and the transcription of the T'ang Annals. 
In my opinion, the transcription *Mwa-lin is intended for the Malindi 
of Edrlsl or Mulanda of Yaqut, now Malindi, south of the Equator, in 
Seyidieh Province of British East Africa. Edrlsl describes this place 
as a large city, the inhabitants of which live by hunting and fishing. 
They salt sea-fish for trade, and also exploit iron-mines, iron being the 
source of their wealth. 6 If this identification be correct, the geographical 
definition of the T'ang Annals (2000 It south-west of Fu-lin) is, of course, 
deficient; but we must not lose sight of the fact that these data rest 
on a hearsay report hailing from Fu-lin, and that, generally speaking, 
Chinese calculations of distances on sea-routes are not to be taken too 
seriously. 7 Under the Ming, the same country appears as jfefc ^ Ma-lin, 
the king of which sent an embassy to China in 141 5 with a gift of 

1 In the transcription hu-man, as given above, followed by the explanation that 
this is the "Persian jujube." The date is not a native of eastern Africa, nor does it 
thrive in the tropics, but it was doubtless introduced there by the Arabs (cf. F. 
Storbeck, Mitt. Sent. Or. Spr., 1914, II, p. 158; A. Engler, Nutzpflanzen Ost- 
Afrikas, p. 12). 

2 Knowledge possessed by the Chinese of the Arabs, p. 25. 

3 China and the Roman Orient, p. 204. 

4 If Mo-lin was on the littoral of the Red Sea, it would certainly be an absurdity 
to define its location as south-west of Jersualem. 

8 Ch. 184, p. 3. 

6 Dozy and de Goeje, Edrlsl's description de 1'Afrique, p. 56 (Leiden, 1866). 

7 Cf. Chinese Clay Figures, pp. 80-81, note. 

390 Sino-Iranica 

giraffes. 1 It likewise appears in the list of countries visited by Cen Ho, 2 
where Ma-lin and La-sa M %& are named, the latter apparently being 
identical with the older Lao-p'o-sa. 3 

The Chinese knew, further, that the date thrives in the country of 
the Arabs (Ta-si), 4 further, in Oman, Basra, and on the Coromandel 
Coast. 5 It is pointed out, further, for Aden and Ormuz. 6 

There is no doubt that the date-palm has existed in southern Persia 
from ancient times, chiefly on the littoral of the Persian Gulf and in 
Mekran, Baluchistan. It is mentioned in several passages of the 
Bundahisn. 7 Its great antiquity in Babylonia also is uncontested 
(Assyrian giUmmaru). 8 Strabo 9 reports how Alexander's army was 
greatly distressed on its march through the barren Gedrosian desert. 
The supplies had to come from a distance, and were scanty and un- 
frequent, so much so that the army suffered greatly from hunger, the 
beasts of burden dropped, and the baggage was abandoned. The army 
was saved by the consumption of dates and the marrow of the palm- 
tree. 10 Again he tells us that many persons were suffocated by eating 
unripe dates. 11 Philostratus speaks of a eunuch who received Apollonius 
of Tyana when he entered the Parthian kingdom, and offered him 
dates of amber color and of exceptional size. 12 In the Province of Fars, 
the date-palm is conspicuous almost everywhere. 13 In Babylon, Persian 
and Aramaic date-palms were distinguished, the former being held in 
greater esteem, as their meat perfectly detaches itself from the stone, 
while it partially adheres in the Aramaic date. 14 The same distinction 

I Ta Miii i t'ufi ii, Ch. 90, p. 24. 
s Min Si, Ch. 304. 

* It is not Ma-lih-la-sa, the name of a single country, as made out by Groene- 
veldt (Notes on the Malay Archipelago, p. 170). 

* T'ai p'in hivan yu ki, Ch. 186, p. 15 b.\ 
5 Hirth, Chau Ju-kua, pp. 133, 137, 96. 

8 Rockhill, T'oung Pao, 1915, p. 609. The word to-Sa-pu, not explained by 
him, represents Arabic duSdb (" date- wine" ; see Leclerc, Traits des simples, 
Vol. II, p. 49). Noldeke (Persische Studien, II, p. 42) explains this word from 
dills ("honey") and Persian db ("water"). 

7 Above, p. 193. 

8 Herodotus, 1, 193; E. Bonavia, Flora of the Assyrian Monuments, p. 3; 
Handcock, Mesopotamian Archaeology, pp. 12-13. 

9 xv, 2, § 7. 

10 Cf. Theophrastus, Histor. plant., IV. iv, 13. 

II Ibid., IV. iv, 5; and Pliny, xm, 9. 

12 C. Joret, Plantes dans l'antiquite\ Vol. II, p. 93. 

13 G. Le Strange, Description of the Province of Fars, pp. 31, 33, 35, 39, 40, 

14 1. Loew, Aramaeische Pflanzennamen, p. 112. 

The Date-Palm 391 

was made in the Sasanian empire: in the tax laws of Khosrau I (a.d. 
531-578), four Persian date-palms were valued and taxed equally with 
six common ones. 1 As already remarked, the Wei and Sui Annals 
attribute the date to Sasanian Persia, and the date is mentioned in 
Pahlavi literature (above, p. 193). At present dates thrive in the low 
plains of Kerman and of the littoral of the Persian Gulf; but the crops 
are insufficient, so that a considerable importation from Bagdad takes 
place. 2 

A. de Candolle 3 asserts, "No Sanskrit name is known, whence it 

j may be inferred that the plantations of the date-palm in western India 

: : are not very ancient. The Indian climate does not suit the species." 

j There is the Sanskrit name kharjura for Phoenix sylvestris, that already 

/occurs in the Yajurveda. 4 This is the wild date or date-sugar palm, 

which is indigenous in many parts of India, being most abundant in 

Bengal, Bihar, on the Coromandel Coast, and in Gujarat. The edible 

[ date (P. dactylifera) is cultivated and self-sown in Sind and the southern 

Panjab, particularly near Multan, Muzaffargarh, the Sind Sagar Doab, 

and in the Trans-Indus territory. It is also grown in the Deccan and 

Gujarat. 5 Its Hindi name is khajilra, Hindustani khajur, from Sanskrit 

kharjura. It is also called sindhi, seindi, sendri, which names allude to 

its origin from Sind. Possibly Sanskrit kharjura and Iranian khurma(n), 

at least as far as the first element is concerned, are anciently related. 

1 N6ldeke, Tabari, p. 245. 

2 Schlimmer, Terminologie, p. 175. 

3 Origin of Cultivated Plants, p. 303. 

4 Macdonell and Keith, Vedic Index, Vol. I, p. 215. 

5 G. Watt, Commercial Products of India, pp. 883, 885. 


36. In regard to the spinach (Spinacia oleracea), Bretschneider 1 
stated that "it is said to come from Persia. The botanists consider 
western Asia as the native country of spinach, and derive the names 
Spinacia, spinage, spinat, epinards, from the spinous seeds; but as the 
Persian name is esfinadsh, our various names would seem more likely 
to be of Persian origin." The problem is not quite so simple, however. 
It is not stated straightforwardly in any Chinese source that the spinach 
comes from Persia; and the name "Persian vegetable" (Pose ts'ai) is 
of recent origin, being first traceable in the Pen ts'ao kan mu, where 
Li Si-cen himself ascribes it to a certain Fan Si-yin 3? :fc HI. 

Strangely enough, we get also in this case a taste of the Can-K'ien 
myth. At least, H. L. Joly 2 asserts, "The Chinese and Japanese Reposi- 
tory says that Chang K'ien brought to China the spinach." The only 
Chinese work in which I am able to find this tradition is the T'un U 
3® iu£, 3 written by Cen Tsiao JIB Wk of the Sung dynasty, who states in 
cold blood that Can K'ien brought spinach over. Not even the Pen 
ts'ao kan mu dares repeat this fantasy. It is plainly devoid of any 
value, in view of the fact that spinach was unknown in the west as 
far back as the second century B.C. Indeed, it was unfamiliar to the 
Semites and to the ancients. It is a cultivation that comes to light 
only in mediaeval times. 

In perfect agreement with this state of affairs, spinach is not men- 
tioned in China earlier than the T'ang period. As regards the literature 
on agriculture, the vegetable makes its first appearance in the Cun §u 
§u W. HJ lif, written toward the end of the eighth century. l Here it is 
stated that the spinach, po-lin H ^ (*pwa-lin), came from the country 
Po-lin St H S3 (*Pwa-lih, Palinga). 

The first Pen ts'ao that speaks of the spinach is the Cen lei pen ts'ao 
written by T'an Sen-wei in a.d. 1108. 5 This Materia Medica describes 
altogether 1746 articles, compared with 11 18 which are treated in the 
Kia yu pu cu pen ts'ao (published in the period Kia-yu, a.d. 1056-64), 
so that 628 new ones were added. These are expressly so designated in 

1 Chinese Recorder, 1871, p. 223. 

2 Legend in Japanese Art, p. 35. 

3 Ch. 75, p. 32 b. 

4 Bretschneider, Bot. Sin., pt. 1, p. 79. 
6 Ch. 29, p. 14 b (print of 1587). 


The Spinach 393 

the table of contents preceding each chapter, and spinach ranks among 
these novelties. Judging from the description here given, it must have 
been a favorite vegetable in the Sung period. It is said to be particularly 
beneficial to the people in the north of China, who feed on meat and 
flour (chiefly in the form of vermicelli), while the southerners, who 
subsist on fish and turtles, cannot eat much of it, because their water 
food makes them cold, and spinach brings about the same effect. 1 
The Kia yii (or hwa) lu £ 3§ (or R) ft by Liu Yu-si ft ^ £§ (a.d. 
772-842) is cited to the effect that "po-lin *<$£ |§ was originally in the 
western countries, and that its seeds came thence to China 2 in the 
same manner as alfalfa and grapes were brought over by Can K'ien. 
Originally it was the country of Po-lin M W., and an error arose in the 
course of the transmission of the word, which is not known to many at 
this time." 

The first and only historical reference to the matter that we have 
occurs in the T'an hui yao, 3 where it is on record, "At the time of the 
Emperor T'ai Tsun (a.d. 627-649), in the twenty-first year of the period 
Ceh-kwan (a.d. 647), Ni-p'o-lo (Nepal) sent to the Court the vegetable 
po-lin fmWt, resembling the flower of the hun-lan %L He (Carthamus 
tinctorius), the fruit being like that of the tsi-li 2^! 1£ (Tribulus ter- 
restris). Well cooked, it makes good eating, and is savory." 4 

This text represents not only the earliest datable mention of the 
vegetable in Chinese records, but in general the earliest reference to it 
that we thus far possess. This document shows that the plant then was 
a novelty not only to the Chinese, but presumably also to the people 
of Nepal; otherwise they would not have thought it worthy of being 
sent as a gift to China, which was made in response to a request of the 

1 John Gerarde (The Herball or Generall Historie of Plantes, p. 260, London, 
*597) remarks, "Spinach is evidently colde and moist, almost in the second degree, 
but rather moist. It is one of the potherbes whose substance is waterie." 

2 According to another reading, a Buddhist monk (seft) is said to have brought 
the seeds over, which sounds rather plausible. G. A. Stuart remarks that the herb 
is extensively used by the monks in their lenten fare. 

3 Ch. 200, p. 14 b (also Ch. 100, p. 3 b). Cf. Ts'efu yuan kwei, Ch. 970, p. 12, 
and Pei hu lu, Ch. 2, p. 19 b (ed. of Lu Sin-yuan). 

* The T'ai p'in yii Ian (Ch. 980, p. 7) attributes this text to the T'ang Annals. 
It is not extant, however, in the account of Nepal inserted in the two T'an Su, nor 
in the notice of Nepal in the T'an hui yao. Pen ts'ao kan mu, T'u Su tsi Ven, and 
Ci wu min Si t'u k'ao (Ch. 5, p. 37) correctly cite the above text from the T'an hui 
yao, with the only variant that the leaves of the po-lin resemble those of the hun- 
lan. The Fun Si wen kien ki (Ch. 7, p. 1 b) by Fun Yen of the ninth century 
(above, p. 232), referring to the same introduction, offers a singular name for the 
spinach in the form $£ $| \ % jS| po-lo-pa-tsao, *pa-la-bat-tsaw, or, if tsao, denot- 
ing several aquatic plants, does not form part of the transcription, *pa-la-bat(bar). 

394 Sino-Iranica 

Emperor T'ai Tsun that all tributary nations should present their 
choicest vegetable products. Yuan Wen j£ 3fc, an author of the Sung 
period, in his work Wen yu kien p'ih §£ M PU W, 1 states that the spinach 
(po-lin) comes from (or is produced in) the country Ni-p'o-lo (Nepal) 
in the Western Regions. 2 The Kia yu pen ts'ao, compiled in a.d. 1057, 
is the first Materia Medica that introduced the spinach into the pharma- 
copoeia. 3 

The colloquial name is po ts'ai ^ ^S ("po vegetable"), po being 
abbreviated for po-lin. According to Wan £i-mou 3: ft 8& (who died 
in 1 591), in his Kwa susu J&WL (Dft, the current name in northern China 
is £*♦ ken ts'ai # $k £& ("red-root vegetable"). The Kwan k'unfan p'u 
uses also the term yih-wu ts'ai ("parrot vegetable"), named for the 
root, which is red, and believed to resemble a parrot. Aside from the 
term Pose ts'ai, the Pen ts'ao kan mu H i* gives the synonymes huh 
ts'ai SI^ ("red vegetable") and yah ^ ts'ai ("foreign vegetable"). 
Another designation is lan-hu ts'ai ("coral vegetable"). 

A rather bad joke is perpetrated by the Min §u ffi #, a description 
of Fu-kien Province written at the end of the sixteenth or beginning of 
the seventeenth century, where the name po-lin is explained as $t It 
po leh ("waves and edges"), because the leaves are shaped like wave- 
patterns and have edges. There is nothing, of course, that the Chinese 
could not etymologize. 5 

There is no account in the traditions of the T'ang and Sung periods 
to the effect that the spinach was derived from Persia; and in view of 
the recent origin of the term "Persian vegetable," which is not even 
explained, we are tempted at the outset to dismiss the theory of 
a Persian origin. Stuart 6 even goes so far as to say that, "as the Chinese 
have a tendency to attribute everything that comes from the south- 
west to Persia, we are not surprised to find this called Pose ts'ao, 'Per- 

1 Ch. 4, p. lib (ed. of Wu yin Hen, 1775). 

2 9i Kk Hi M W. % 3 $£ 13- This could be translated also, "in the 
Western Regions and in the country Ni-p'o-lo." 

3 Ci wu min U t'u k'ao, Ch. 4, p. 38 b. 

« Ch. 8, p. 87 b. 

6 Of greater interest is the following fact recorded in the same book. The 
spinach in the north of China is styled "bamboo (cu ^f) po-lin," with long and 
bitter stems; that of Fu-kien is termed "stone (U ^3) po-lin," and has short and 
sweet stems. — The Min $u, in 154 chapters, was written by Ho K'iao-yuan -fpT ll? 
jH from Tsin-kian in Fu-kien; he obtained the degree of tsin U in 1586 (cf. Cat. of 
the Imperial Library, Ch. 74, p. 19). 

8 Chinese Materia Medica, p. 417. 

The Spinach 395 

sian vegetable.' ' n There is, however, another side to the case. In all 
probability, as shown by A. de Candolle, 2 it was Persia where the 
spinach was first raised as a vegetable; but the date given by him, 
"from the time of the Graeco-Roman civilization," is far too early.' 
A. deCandolle's statement that the Arabs did not carry the plant to Spain 
has already been rectified by L. Leclerc; 4 as his work is usually not in 
the hands of botanists or other students using de Candolle, this may 
aptly be pointed out here. 

According to a treatise on agriculture {Kitab el-jalaha) written by 
Ibn al-Awwam of Spain toward the end of the eleventh century, spinach 
was cultivated in Spain at that time. 6 Ibn Haddjaj had then even 
written a special treatise on the cultivation of the vegetable, saying that 
it was sown at Sevilla in January. From Spain it spread to the rest of 
Europe. Additional evidence is afforded by the very name of the 
plant, which is of Persian origin, and was carried by the Arabs to Europe. 
The Persian designation is aspanah, aspandj or asfindj; Arabic isfenah 
or isbenah. Hence Mediaeval Latin spinachium or spinarium, 6 Spanish 

1 The outcry of Waiters (Essays on the Chinese Language, p. 347) against the 
looseness of the term Po-se, and his denunciation of the "Persian vegetable" as "an 
example of the loose way in which the word is used," are entirely out of place. It 
is utterly incorrect to say that "they have made it include, beside Persia itself, Syria, 
Turkey, and the Roman Empire, and sometimes they seem to use it as a sort of 
general designation for the abode of any barbarian people to the south-west of 
the Middle Kingdom." Po-se is a good transcription of Parsa, the native designa- 
tion of Persia, and strictly refers to Persia and to nought else. When P. P. Smith applied 
the name po-ts'ai to Convolvulus reptans, this was one of the numerous confusions 
and errors to which he fell victim. Likewise is it untrue, as asserted by Watters, 
that the term has been applied even to beet and carrot and other vegetables not 
indigenous in Persia. As on so many other points, Watters was badly informed on 
this subject also. 

3 Origin of Cultivated Plants, pp. 98-100. 

* This conclusion, again, is the immediate outcome of Bretschneider's Chang- 
kienomania: for A. de Candolle says, " Bretschneider tells us that the Chinese 
name signifies 'herb of Persia,' and that Western vegetables were commonly intro- 
duced into China a century before the Christian era." 

♦Traite" des simples, Vol. I, p. 61. 

6 L. Leclerc, Histoire de la mddecine arabe, Vol. II, p. 112. The Arabic work 
has been translated into French by Clement-Mullet under the title Ibn al Awwam, 
le livre de l'agriculture (2 vols., Paris, 1864-67). De Candolle's erroneous theory 
that "the European cultivation must have come from the East about the fifteenth 
century," unfortunately still holds sway, and is perpetuated, for instance, in the 
last edition of the Encyclopaedia Britannica. 

* The earliest occurrence of this term quoted by Du Cange refers to the year 
135 1, and is contained in the Transactio inter Abbatem et Monachos Crassenses. 
Spinach served the Christian monks of Europe as well as the Buddhists of China. 
O. Schrader (Reallexikon, p. 788) asserts that the vegetable is first mentioned by 
Albertus Magnus (1193-1280) under the name spinachium, but he fails to give a 

396 Sino-Iranica 

espinaca, Portuguese espinafre or espinacio, Italian spinace or spinaccio, 
Provencal espinarc, Old French espinoche or epinoche, French epinard. 1 
The Persian word was further adopted into Armenian spanax or 
asbanax, Turkish spandk or ispandk, Comanian yspanac, Middle 
Greek spinakion, Neo-Greek spanaki{pn) or spanakia (plural). 
There are various spellings in older English, like spynnage, 
spenege, spinnage, spinage, etc. In English literature it is not men- 
tioned earlier than the sixteenth century. W. Turner, in his 
"Herball" of 1568, speaks of "spinage or spinech as an herbe lately 
found and not long in use." 

However, in the latter part of the sixteenth century, spinach was 
well known and generally eaten in England. D. Rembert Dodoens 2 
describes it as a perfectly known subject, and so does John Gerarde, 3 
who does not even intimate that it came but recently into use. The 
names employed by them are Spanachea, Spinachia, Spinacheum olus, 
Hispanicum olus, English spinage and spinach. John Parkinson 4 
likewise gives a full description and recipes for the preparation of the 

The earliest Persian mention of the spinach, as far as I know, is 
made in the pharmacopoeia of Abu Mansur. 5 The oldest source cited 
by Ibn al-Baitar (1197-1248) 6 on the subject is the "Book of Nabathaean 
Agriculture" (Falaha nabatiya), which pretends to be the Arabic trans- 
lation of an ancient Nabathasan source, and is believed to be a forgery 
of the tenth century. This book speaks of the spinach as a known 
vegetable and as the most harmless of all vegetables; but the most 
interesting remark is that there is a wild species resembling the culti- 
vated one, save that it is more slender and thinner, that the leaves are 

specific reference. It is a gratuitous theory of his that the spinach must have been 
brought to Europe by the Crusaders; the Arabic importation into Spain has escaped 
him entirely. 

1 The former derivation of the word from "Spain" or from spina ("thorn"), in 
allusion to the prickly seeds, moves on the same high level as the performance of the 
Min $u. Littre" cites M£nagier of the sixteenth century to the effect, "Les espinars 
sont ainsi appell6s a cause de leur graine qui est espineuse, bien qu'il y en ait de ronde 
sans piqueron." In the Supplement, Littr6 points out the oriental origin of the word, 
as established by Devic. 

2 A Niewe Herball, or Historie of Plants, translated by H. Lyte, p. 556 (Lon- 
don, 1578). 

* The Herball or Generall Historie of Plantes, p. 260 (London, 1597). 

* Paradisus in sole paradisus terrestris, p. 496 (London, 1629). 
5 Achundow, Abu Mansur, p. 6. 

8 L. Leclerc, Traits des simples, Vol. I, p. 60. 

The Spinach 397 

more deeply divided, and that it rises less from the ground. 1 A. de 
Candolle states that "spinach has not yet been found in a wild state, 
unless it be a cultivated modification of Spinacia tetandra Steven, which 
is wild to the south of the Caucasus, in Turkistan, in Persia, and in 
Afghanistan, and which is used as a vegetable under the name of 
Samum." The latter word is apparently a bad spelling or misreading 
for Persian lomln or iwraln (Armenian zomin and Somin), another 
designation for the spinach. 

The spinach is not known in India except as an introduction by the 
English. The agriculturists of India classify spinach among the English 
vegetables. 2 The species Spinacia tetrandra Roxb., for which Rox- 
burgh 3 gives the common Persian and Arabic name for the spinach, 
and of which he says that it is much cultivated in Bengal and the 
adjoining provinces, being a pot-herb held in considerable estimation 
by the natives, may possibly have been introduced by the Moham- 
medans. As a matter of fact, spinach is a vegetable of the temperate 
zones and alien to tropical regions. A genuine Sanskrit word for the 
spinach is unknown. 4 Nevertheless Chinese po-lin, *pwa-lin, must 
represent the transcription of some Indian vernacular name. In Hin- 
dustani we have palak as designation for the spinach, and palan or 
palak as name for Beta vulgaris, Pustu palak ? apparently developed 
from Sanskrit palahka, palankya, palakyu, palakya, to which our 
dictionaries attribute the meaning "a kind of vegetable, a kind of 
beet-root, Beta bengalensis"; in Bengali palun.*' To render the coin- 
cidence with the Chinese form complete, there is also Sanskrit Palakka 

1 Perhaps related to A triplex L., the so-called wild spinach, chiefly cultivated 
in France and eaten like spinach. The above description, of course, must 
not be construed to mean that the cultivated spinach is derived from the 
so-called wild spinach of the Nabathaeans. The two plants may not be in- 
terrelated at all. 

1 N. G. Mukerji, Handbook of Indian Agriculture, 2d ed., p. 300 (Calcutta, 
1907); but it is incorrect to state that spinach originally came from northern Asia. 
A. de Candolle (op. cit., p. 99) has already observed, "Some popular works repeat 
that spinach is a native of northern Asia, but there is nothing to confirm this sup- 

'Flora Indica, p. 718. 

* A. Borooah, in his English-Sanskrit Dictionary, gives a word cakaprabheda 
with this meaning, but this simply signifies "a kind of vegetable," and is accord- 
ingly an explanation. 

6 H. W. Bellew, Report on the Yusufzais, p. 255 (Lahore, 1864). 

6 Beta is much cultivated by the natives of Bengal, the leaves being consumed 
in stews (W. Roxburgh, Flora Indica, p. 260). Another species, Beta maritima, is 
also known as "wild spinach." It should be remembered that the genus Beta belongs 
to the same family (Chenopodiaceae) as Spinacia. 

398 Sino-Iranica 

or Palaka 1 as the name of a country, which has evidently resulted in 
the assertion of Buddhist monks that the spinach must come from a 
country Palinga. The Nepalese, accordingly, applied a word relative 
to a native plant to the newly-introduced spinach, and, together with 
the product, handed this word on to China. The Tibetans never became 
acquainted with the plant; the word spo ts'od, given in the Polyglot 
Dictionary, 2 is artificially modelled after the Chinese term, spo (pro- 
nounced po) transcribing Chinese po, and ts'od meaning "vegetable." 

Due regard being paid to all facts botanical and historical, we are 
compelled to admit that the spinach was introduced into Nepal from 
some Iranian region, and thence transmitted to China in a.d. 647. 
It must further be admitted that the Chinese designation "Persian 
vegetable," despite its comparatively recent date, cannot be wholly 
fictitious, but has some foundation in fact. Either in the Yuan or in 
the Ming period (more probably in the former) the Chinese seem to 
have learned the fact that Persia is the land of the spinach. I trust that 
a text to this effect will be discovered in the future. All available his- 
torical data point to the conclusion that the Persian cultivation can 
be but of comparatively recent origin, and is not older than the sixth 
century or so. The Chinese notice referring it to the seventh century 
is the oldest in existence. Then follow the Nabathaean Book of Agri- 
culture of the tenth century and the Arabic introduction into Spain 
during the eleventh. 

1 The latter form is noted in the catalogue of the Mahamayari, edited by S. 
L£vi (Journal asiatique, 1915, I, p. 42). 
J Ch. 27, p. 19 b. 


37. In the preceding notes we observed that the name for a species 
of Beta was transferred to the spinach in India and still serves in China 
as designation for this vegetable. We have also a Sino-Iranian name 
for a Beta, J? H, kiin-Va, *gwun-d'ar, which belonged to the choice 
vegetables of the country ^ Iffc Mo-lu, *Mar-luk, in Arabia. 1 The 
Cen su wen Wi f& 3&C 2 says that it is now erroneously called ken ta ts'ai 
$1 ^C £(k or ta ken ts'ai, which is identical with tien ts'ai Stt $& ("sweet 
vegetable "). Stuart 5 gives the latter name together with M H kiin-Va, 
identifying it with Beta vulgaris, the white sugar beet, which he says 
grows in China. Stuart, however, is mistaken in saying that this plant 
is not mentioned in the Pen ts'ao. It is noted both in the Cen lei pen 
ts'ao* and the Pen ts'ao kan mu, h the latter giving also the term kiin-Va, 
which is lacking in the former work. Li Si-cen observes with reference 
to this term that its meaning is unexplained, a comment which usually 
betrays the foreign character of the word, but he fails to state the 
source from which he derived it. There is no doubt that this kiin-Va 
is merely a graphic variant of the above % ||. The writing M is as 
early as the T'ang period, and occurs in the Yu yan tsa tsu, 6 where the 
leaves of the yu tien ts'ao $i S W- ("herb with oily spots") are com- 
pared to those of the kiin-Va. 7 A description of the kiin-Va is not con- 
tained in that work, but from this incidental reference it must be 
inferred that the plant was well known in the latter half of the ninth 

Beta vulgaris is called in New Persian lugundur or legonder, and 
Is mentioned by Abu Mansur. 8 The corresponding Arabic word is 
silk. 9 The Chinese transcription made in the T'ang period is apparently 
based on a Middle-Persian form of the type *gundar or *gundur. Beta 
vulgaris is a Mediterranean and West-Asiatic plant grown as far as the 

1 T'ai p'iA hwan yu ki, Ch. 186, p. 16 b. 

1 Ch. 12, p. 3. This work was published in 1884 by Ho Yi-hin $5 fj£ fjf , 

* Chinese Materia Medica, p. 68. 

* Ch. 28, p. 9. 

6 Ch. 27, p. 1 b. Cf. also Yatnato honzd, Ch. 5, p. 26. 

6 Ch. 9, p. 9 b. 

7 "On each leaf there are black spots opposite one another." 

8 Achundow, Abu Mansur, p. 81. 

9 Leclerc, Traits des simples, Vol. II, p. 274. 


400 Sino-Iranica 

Caspian Sea and Persia. According to de Candolle, 1 its cultivation 
does not date from more than three or four centuries before our era. 
The Egyptian illustration brought forward by F. Woenig 2 in favor of 
the assumption of an early cultivation in Egypt is not convincing to 

It is therefore probable, although we have no record referring to the 
introduction, that Beta vulgaris was introduced into China in the T'ang 
period, perhaps by the Arabs, who themselves brought many Persian 
words and products to China. For this reason Chinese records some- 
times credit Persian words to the Ta-si (Arabs); for instance, the 
numbers on dice, which go as Ta-si, but in fact are Persian. 3 

The real Chinese name of the plant is tien ts'ai ^ ?£., the first 
character being explained in sound and meaning by Stt tien ("sweet"). 
Li Si-cen identifies tien ts'ai with kiln-fa. The earliest description 
of tien ts'ai comes from Su Kuh of the T'ang, who compares its leaves to 
those of hn ma Tt i^ (Actea spicdta, a ranunculaceous plant), adding 
that the southerners steam the sprouts and eat them, the dish being very 
fragrant and fine. 4 It is not stated, however, that tien ts'ai is an im- 
ported article. 

38. Reference was made above to the memorable text of the T'an 
hui yao, in which are enumerated the vegetable products of foreign 
countries sent to the Emperor T'ai Tsuh of the T'ang dynasty at his 
special request in a.d. 647. After mentioning the spinach of Nepal, 
the text continues thus: — 

"Further, there was the ts'o ts'ai HJfeZH ('wine vegetable') with 
broad and long leaves. 5 It has a taste like a good wine and k'u ts'ai 
^ ^ ('bitter vegetable,' lettuce, Lactuca), and in its appearance is like 
kii J?, 6 but its leaves are longer and broader. Although it is somewhat 
bitter of taste, eating it for a long time is beneficial. Hu k'in $J Jr 

1 Origin of Cultivated Plants, p. 59; see also his G6ographie botanique, p. 831 

2 Pflanzen im alten Aegypten, p. 218. 
8 See T'oung Pao, Vol. I, 1890, p. 95. 

4 A tien ts'ai mentioned by T'ao Hun-kin, as quoted in the Pen ts'ao kan mu, 
and made into a condiment la Rp for cooking-purposes, is apparently a different 

6 The corresponding text of the Ts'e fu yuan kwei (Ch. 970, p. 12) has the 
addition, "resembling the leaves of the Sen-hivo ^ }K-" The text of the Pei hu 
lu (Ch. 2, p. 19 b) has, "resembling in its appearance the Sen-hwo, but with leaves 
broader and longer." This tree, also called kin t'ien j§£ ^ (see Yu yan tsa tsu, 
Ch. 19, p. 6), is believed to protect houses from fire; it is identified with Sedum erythro- 
stictum or Sempervivum tectorum (Bretschneider, Bot. Sin., pt. Ill, No. 205; 
Stuart, Chinese Materia Medica, p. 401). 

6 A general term for plants like Lactuca, Cichorium, Sonchus. 

Sugar Beet and Lettuce 401 

resembles in its appearance the k'in ?? ('celery,' Apium graveolens), 
and has a fragrant flavor." 

Judging from the description, the vegetable ts*o ts'ai appears to have 
been a species of Lactuca, Cidwrium, or Sonchus. These genera are 
closely allied, belonging to the family Cichoraceae, and are confounded 
by the Chinese under a large number of terms. A. de Candolle 1 
supposed that lettuce (Lactuca sativa) was hardly known in China at 
an early date, as, according to Loureiro, Europeans had introduced it 
into Macao. 2 With reference to this passage, Bretschneider 3 thinks 
that de Candolle "may be right, although the Pen ts'ao says nothing 
about the introduction; the Uh ts'ai ^ ^S (the common name of lettuce 
at Peking) or pai-kii & H seems not to be mentioned earlier than by 
writers of the T'ang (618-906)." Again, de Candolle seized on this 
passage, and embodied it in his "Origin of Cultivated Plants" (p. 96). 
The problem, however, is not so simple. Bretschneider must have 
read the Pen ts'ao at that time rather superficially, for some species of 
Lactuca is directly designated there as being of foreign origin. Again, 
twenty-five years later, he wrote a notice on the same subject, 4 in which 
not a word is said about foreign introduction, and from which, on the 
contrary, it would appear that Lactuca, Cichorium, and Sonchus, have 
been indigenous to China from ancient times, as the bitter vegetable 
(k'u ts'ai) is already mentioned in the Pen kin and Pie lu. The terms 
pai kil Q H and k'u kii i^r H are supposed to represent Cichorium 
endivia; and wo-kii jaU g, Lactuca sativa. In explanation of the latter 
name, Li Si-cen cites the Mo k'o hui si M $r W JP by P'en C'eh 1£ 3$, 
who wrote in the first half of the eleventh century, as saying that wo 
ts'ai i?S ^ ("wo vegetable") came from the country iBt Kwa, and hence 
received its name. 5 The TsHn i lu W j^ $fc, a work by T"ao Ku P^ Wi 
of the Sung period, says that "envoys from the country Kwa came 
to China, and at the request of the people distributed seeds of a vegetable ; 
they were so generously rewarded that it was called ts'ien kin ts'ai 
^ 4£ #5 ('vegetable of a thousand gold pieces'); now it is styled wo- 

1 Geographie botanique, p. 843. 

2 This certainly is a weak argument. The evidence, in fact, proves nothing. 
Europeans also introduce their own sugar and many other products of which China 
has a great plenty. 

8 Chinese Recorder, 1871, p. 223. 

* Bot. Sin., pt. Ill, No. 257. 

5 1 do not know how Stuart (p. 229) gets at the definition "in the time of the 
Han dynasty." The same text is also contained in the Sii po wu li (Ch. 7, p. 1 b), 
written by Li Si ^ ^ about the middle of the twelfth century. 

402 Sino-Iranica 

kit." 1 These are vague and puerile anecdotes, without chronological 
specification. There is no country Kwa, which is merely distilled from 
the character jK&, and no such tradition appears in any historical text. 2 
The term wo-kii was well known under the T'ang, being mentioned in 
the Pen ts'ao H i of C'en Ts'an-k'i, who distinguishes a white and a 
purple variety, but is silent as to the point of introduction. 3 This 
author, however, as can be shown by numerous instances, had a keen 
sense of foreign plants and products, and never failed to indicate them 
as such. There is no evidence for the supposition that Lactuca was 
introduced into China from abroad. All there is to it amounts to this, 
that, as shown by the above passage of the T'an hui yao, possibly supe- 
rior varieties of the West were introduced. 

In Persia, Lactuca sativa (Persian kdhu) occurs both wild and culti- 
vated. 4 Cichoreum is kasnl in Persian, hindubd in Arabic and Osmanli. 5 

39. The hu k'in, mentioned in the above text of the T'an hui yao, 
possibly represents the garden celery, Apium graveolens (Persian kerefs 
or karqfs) (or possibly parsley, Apium petroselinum) of the west. 6 It 
appears to be a different plant from the hu k'in mentioned above (p. 196). 

Hu k'in is likewise mentioned among the best vegetables of the 
country ^c ^ Mo-lu, *Mwat-luk, Mar-luk, in Arabia. 7 

In order to conclude the series of vegetables enumerated in the 
text of the T'an hui yao, the following may be added here. 

In a.d. 647 the king of Gandhara (in north-western India) sent to 
the Chinese Court a vegetable styled ju-t'u ^ i ££ (" Buddha-land 
vegetable"), each stem possessing five leaves, with red flowers, a yellow 
pith, and purple stamens. 8 

1 I have looked up the text of the Ts'in i lu, which is reprinted in the T'an Sun 
ts'un $u and Si yin hiian ts'un $u. The passage in question is in Ch. 2, p. 7 b, and 
printed in the same manner as in the Pen ts'ao ban tnu, save that the country is called 
Kao ]lb, not Kwa fSf. It is easy to see that these two characters could be con- 
founded, and that only one of the two can be correct; but Kao does not help us any 
more than Kwa. Either name is fictitious as that of a country. 

2 We have had several other examples of alleged names of countries being 
distilled out of botanical names. 

s K'ou Tsuh-§i is likewise; see his Pen ts'ao yen i (Ch. 19, p. 2). 
* Schlimmer, Terminologie, p. 337. 

5 See Achundow, Abu Mansur, p. 146; E. Seidel, Mechithar, p. 134; Leclerc, 
TraitS des simples, Vol. II, p. 28. 

6 Cf. Achundow, Abu Mansur, pp. no, 257. Celery is cultivated only in a few 
gardens of Teheran, but it grows spontaneously and abundantly in the mountains 
of the Bakhtiaris (Schlimmer, Terminologie, p. 43). 

7 T'ai p'in hwan yii ki, Ch. 186, p. 16 b. 

8 T'an hui yao, Ch. 200, p. 4 b; and T'an $u, Ch. 221 B, p. 7. The name of 
Gandhara is abbreviated into *d'ar, but in the corresponding passage of the T'an 
hui yao (Ch. 100, p. 3 b) and in the Ts'e fu yuan kwei (Ch. 970, p. 12) the name is 
written completely ||| ^ Kien-ta, *G'an-d'ar. 


40. In regard to Ricinus communis (family Euphorbiaceae) the 
accounts of the Chinese are strikingly deficient and unsatisfactory. 
There can be no doubt that it is an introduced plant in China, as it 
occurs there only in the cultivated state, and is not mentioned earlier 
than the T'ang period (618-906) with an allusion to the Hu. 1 Su Kun 
states in the T'an pen ts'ao, "The leaves of this plant which is culti- 
vated by man resemble those of the hemp (Cannabis sativa), being very 
large. The seeds look like cattle-ticks (niu pei 4 1 4$) . 2 The stems of 
that kind which at present comes from the Hu 3 are red and over ten 
feet high. They are of the size of a tsao kia -S $S (Gleditschia sinensis). 
The kernels are the part used, and they are excellent." It would seem 
from this report that two kinds of Ricinus are assumed, one presumably 
the white-stemmed variety known prior to Su Kuh's time, and the red- 
stemmed variety introduced in his age. Unfortunately we receive no 
information as to the exact date and provenience of the introduction. 

The earliest mention of the plant is made by Herodotus, 4 who 
ascribes it to the Egyptians who live in the marshes and use the oil 
pressed from the seeds for anointing their bodies. He calls the plant 
silliky prion* and gives the Egyptian name as kiki. 6 In Hellas it grows 
spontaneously (avrdnara <£v€tcu), but the iEgyptians cultivate it along 
the banks of the rivers and by the sides of the lakes, where it produces 
fruit in abundance, which, however, is malodorous. This fruit is 

1 Pen ts'ao kan mu, Ch. 17 A, p. II. Bretschneider (Chinese Recorder, 1871, 
p. 242) says that it cannot be decided from Chinese books whether Ricinus is in- 
digenous to China or not, and that the plant is not mentioned before the T'ang. 
The allusion to the Hu escaped him. 

2 Hence the name j£ or j[ffc £| P 6 * ma (only in the written language) for the 
plant (Peking colloquial ta ma, "great hemp ")• This etymology has already been ad- 
vanced by Su Sun of the Sung and confirmed by Li Si-cen, who explains the insect as 
the "louse of cattle." This interpretation appears to be correct, for it represents a 
counterpart to Latin ricinus, which means a "tick": Nostri earn ricinum vocant a 
similitudine seminis (Pliny, xv, 7, § 25). The Chinese may have hit upon this simile 
independently, or, what is even more likely, received it with the plant from the West. 

1 This appears to be the foundation for Stuart's statement (Chinese Materia 
Medica, p. 378) that the plant was introduced from "Tartary." 

4 n, 94. 

s The common name was tcpbrwr (Theophrastus, Hist, plant., I. x, 1), Latin 

5 This word has not yet been traced in the hieroglyphic texts, but in Coptic. 
In the demotic documents Ricinus is deqam (V. Loret, Flore pharaonique, p. 49). 


404 Sino-Iranica 

gathered, and either pounded and pressed or roasted and boiled, and 
the oily fluid is collected. It is found to be unctuous and not inferior to 
olive-oil for burning in lamps, save that it emits a disagreeable odor. 
Seeds of Ricinus are known from Egyptian tombs, and the plant is still 
cultivated in Egypt. Pliny 1 states that it is not so long ago that the 
plant was introduced into Italy. A. de Candolle 2 traces its home to 
tropical Africa, and I agree with this view. Moreover, I hold that it was 
transplanted from Egypt to India, although, of course, we have no 
documentary proof to this effect. Ricinus does not belong to the plants 
which were equally known to the Iranians and Indo-Aryans. It is not 
mentioned in the Vedas or in the Laws of Manu. 3 The first datable 
references to it occur in the Bower Manuscript, where its oil and root 
are pointed out under the names eranda, gandharva, rubugaka, and 
vaksana. Other names are ruvu, ruvuka, or ruvilka, citraka, gandharva- 
hastaka, vyaghrapuccha (" tiger 's-tail"). The word eranda has become 
known to the Chinese in the form i-lan & Hi, 4 and was adopted into the 
language of Kuca (Tokharian B) in the form hiranda. 5 From India 
the plant seems to have spread to the Archipelago and Indo-China 
(Malayan, Sunda, and Javanese jarak; Khmer lohon; Annamese du du 
tran, kai-dua, or kai-du-du-tia; Cam tamnon, lahaun, lahon). 6 The 
Miao and the Lo-lo appear to be familiar with the plant: the former 
call it zrwa-fio; 7 the latter, l K e-tu-ma (that is, "fruit for the poisoning 
of dogs"). 8 

In Iran the cultivation of Ricinus has assumed great importance, 
but no document informs us as to the time of its transplantation. It 
may be admitted, however, that it was well known there prior to our 
era. 9 The Persian name is bedanjir, pandu, punde, or pendu; in Arabic 
it is xarva or xirva. 



1 xv, 7, § 25. 

2 Origin of Cultivated Plants, p. 422. 

3 Joret, Plantes dans l'antiquit6, Vol. II, p. 270. 
* Fan yi tnin yi tsi, section 24. 

5 S. Levi, Journal asiatique, 191 1, II, p. 123. 

6 On the cultivation in Indo-China, see Perrot and Hurrier, Mat. m£d. et 
pharmacop^e sino-annamites, p. 107. Regarding the Archipelago, see A. de Can- 
dolle, op. tit., p. 422; W. Marsden, History of Sumatra, p. 92; J. Crawfurd, 
History of the Indian Archipelago, Vol. I, p. 382. The plant is reported wild from 
Sumatra and the Philippines, but the common Malayan name jarak hints at an 
historical distribution. 

7 F. M. Savina, Dictionnaire miao-tseu-francais, pp. 205, 235. 

8 P. Vial, Dictionnaire francais-lolo, p. 290. Also the Arabs used Ricinus as a 
dog-poison (Leclerc, Traits des simples, Vol. II, p. 20). 

9 Joret, op. tit., p. 72. 


41. Iran was the centre from which the almond (Amygdalus com- 
munis or Prunus amygdalus) spread, on the one hand to Europe, and on 
the other to China, Tibet, and India. As to India, it is cultivated but 
occasionally in Kashmir and the Panjab, where its fruits are mediocre. 
It was doubtless imported there from Iran. The almond yields a gum 
which is still exported from Persia to Bombay, and thence re-exported 
to Europe. 1 The almond grows spontaneously in Afghanistan and 
farther to the north-east in the upper Zarafshan valley, and in the 
Chotkal mountains at an altitude of 1 000-1300 m, also in Aderbeidjan, 
Kurdistan, and Mesopotamia. According to Schlimmer, 2 Amygdalus 
coparia is very general on the high mountains, and its timber yields 
the best charcoal. 3 

The Greeks derived the almond from Asia Minor, and from Greece 
it was apparently introduced into Italy. 4 In the northern part of Media, 
the people subsisted upon the produce of trees, making cakes of apples, 
sliced and dried, and bread of roasted almonds. 5 A certain quantity of 
dried sweet almonds was to be furnished daily for the table of the 
Persian kings. 6 The fruit is mentioned in Pahlavi literature (above, 

P- 193)- 

The Yin yai hn Ian mentions almonds among the fruit grown in 
Aden. 7 The Arabic name is lewze or lauz. Under this name the medicinal 
properties of the fruit are discussed in the Persian pharmacopoeia of 
Abu Mansur, who knew both the sweet almond (bdddm-i Hrln) and the 
bitter one (bddam-i talx)* It is curious that bitter almonds were used 
as currency in the empire of the Moguls. They were brought into the 

1 G. Watt, Commercial Products of India, p. 905; and Dictionary, Vol. VI, 
P- 343- Joret, Plantes dans l'antiquite\ Vol. II, p. 279. W. Roxburgh (Flora 
Indica, p. 403) concluded that the almond is a native of Persia and Arabia, whereas 
it does not succeed in India, requiring much nursing to keep it alive. 

2 Terminologie, p. 33. 

3 A really wild almond is said to be very common in Palestine and Syria (A. 
Aaronsohn, Agric. and Bot. Explorations in Palestine, p. 14). 

4 Hehn, Kulturpflanzen, pp. 393, 402; Fluckiger and Hanbury, Pharma- 
cographia, pp. 244, 245. 

6 Strabo, XI. xiii, 11. 

6 Polyaenus, Strategica, iv, 32. 

7 Rockhill, T'oung Pao, 1915, p. 609. 

8 Achundow, Abu Mansur, p. 128. 


406 Sino-Iranica 

province of Gujarat from Persia, where they grow in dry and arid 
places between rocks; they are as bitter as colocynth, and there is no 
fear that children will amuse themselves by eating them. 1 

What Watters 2 has stated about the almond is for the greater part 
inexact or erroneous. "For the almond which does not grow in China 
the native authors and others have apparently only the Persian name 
which is Badan. This the Chinese transcribe pa-tan A $1 or EL JL and 
perhaps also, as suggested by Bretschneider, pa-lan IE $f ." First, the 
Persian name for the almond is badam; second, the Chinese characters 
given by Watters are not apt to transcribe this word, as the former 
series answers to ancient *pat-dam, the latter to *pa-dan. Both A 
and EL only had an initial labial surd, but never a labial sonant, and 
for this reason could not have been chosen for the transcription of a 
foreign ba in the T'ang period, when the name of the almond made its 
d£but in China. Further, the character EL, which was not possessed 
of a final labial nasal, would make a rather bad reproduction of the 
required element dam. In fact, the characters given by Watters are 
derived from the Pen ts'ao kan tnu, 3 and represent merely a comparative- 
ly modern readjustment of the original form made at a time when 
the transposition of sonants into surds had taken effect. The first form 
given by Watters, as stated in the Pen ts % ao itself, is taken from the 
Yin San Zen yao (see p. 236), written by Ho Se-hwi during the Yuan 
period; while the second form is the work of Li Si-Sen, as admitted by 
himself, and accordingly has no phonetic value whatever. 4 Indeed, we 
have a phonetically exact transcription of the Iranian term, handed 
down from the T'ang period, when the Chinese still enjoyed the pos- 
session of a well-trained ear, and, in view of the greater wealth of sounds 
then prevailing in their speech, also had the faculty of reproducing 
them with a fair degree of precision. This transcription is presented by 
§1 $£ p'o-tan, *bwa-dam, almond (Amygdalus communis or Prunus 
amygdalus), which actually reproduces Middle Persian vadam, New 
Persian badam (Kurd badem, be'iv and baif, "almond-tree")- 5 This term, 

1 Ta vernier, Travels in India, Vol. I, p. 27. 

2 Essays on the Chinese Language, p. 348. 

* Ch. 29, p. 4. Hence adopted also by the Japanese botanists (Matsumura, 
No. 2567), but read amendo (imitation of our word). 

4 He further gives as name for the almond hu-lu-ma %£. $H M. = Persian xurma 
(khurma), but this word properly refers to the date (p. 385). From the Ta Min i 
t'un li (Ch. 89, p. 24), where the almonds of Herat are mentioned, it appears that 
hu-lu-ma (xurma) was the designation of a special variety of almond, "resembling 
a jujube and being sweet." 

6 The assertion of Stuart (Chinese Materia Medica, p. 40), that pa-tan may refer 
to some country in Asia Minor or possibly be another name for Persia, is erroneous. 

The Almond 407 

as far as I know, is first mentioned in the Yu yah tsa tsu, 1 where it is 
said, "The flat peach fin !# grows in the country Po-se (Persia), where 
it is styled p'o-tan. The tree reaches a height of from fifty to sixty feet, 
and has a circumference of four or five feet. Its leaves resemble those 
of the peach, but are broader and larger. The blossoms, which are 
white in color, appear in the third month. When the blossoms drop, the 
formation of the fruit has the appearance of a peach, but the shape 
is flat. Hence they are called 'flat peaches.' The meat is bitter and 
acrid, and cannot be chewed; the interior of the kernel, however, is 
sweet, and is highly prized in the Western Regions and all other coun- 
tries." Although the fact of the introduction of the plant into China 
is not insisted upon by the author, Twan C'en-si, his description, which 
is apparently based on actual observation, may testify to a cultivation 
in the soil of his country. This impression is corroborated by the testi- 
mony of the Arabic merchant Soleiman, who wrote in a.d. 851, and 
enumerates almonds among the fruit growing in China. 2 The cor- 
rectness of the Chinese reproduction of the Iranian name is confirmed 
by the Tibetan form ba-dam, Uigur and Osmanli badam, and Sanskrit 
vatdma or bdddtna, derived from the Middle Persian. 3 

The fundamental text of the Yu yah tsa tsu has unfortunately es- 
caped Li Si-cen, author of the Pen ts'ao kah tnu, and he is accordingly 
led to the vague definition that the almond comes from the old terri- 
tory of the Mohammedans; in his time, he continues, the tree occurred 
in all places West of the Pass (Kwan si; that is, Kan-su and Sen-si). 
The latter statement is suppressed in Bretschneider's translation of 
the text, 4 probably because it did not suit his peremptory opinion that 
the almond-tree does not occur in China. He did not know, either, of 
the text of the Yu yah tsa tsu, and his vague data were adopted by A. 
de Candolle. 8 

Loureiro 6 states that the almond is both wild and cultivated in 

1 Ch. 18, p. 10 b. 

2 M. Reinaud, Relation des voyages, Vol. I, p. 22. 

* Cf. the writer's Loan-Words in Tibetan, No. m. It should be repeated also 
in this place that the Tibetan term p'a-tin, which only means "dried apricots," 
bears no relation to the Persian designation of the almond, as wrongly asserted by 
Watters. — The almond is also known to the Lo-lo (Nyi Lo-lo ni-ma, Ahi Lo-lo 
i-ni-zo, i-sa). 

* Chinese Recorder, 1870, p. 176. 

s Origin of Cultivated Plants, p. 219. He speaks erroneously of the Pen ts'ao 
published in the tenth or eleventh century. Bretschneider, of course, meant the 
Pen ts'ao of the sixteenth century. 

6 Flora cochinchinensis, p. 316. Perrot and Hurrier (Matiere m^dicale et 
pharm. sino-annamites, p. 153) have an Amygdalus cochinchinensis for Annam. 

408 Sino-Iranica 

China. Bunge says that it is commonly cultivated in North China; but 
that recent botanists have not seen it in South China, and the one 
cultivated near Peking is Prunus davidiana, a variety of P. persica. 1 
These data, however, are not in harmony with Chinese accounts which 
attribute the cultivation of the almond to China; and it hardly sounds 
plausible that the Chinese should confound with this tree the apricot, 
which has been a native of their country from time immemorial. 
Watters asserts that "the Chinese have mixed up the foreign almond 
with their native apricot. The name of the latter is kin ■&, and the 
kernels of its fruit, when dried for food, are called hin-Zen 1*? C; This 
name is given also to the kernels of almonds as imported into China 
from their resemblance in appearance and to some extent in taste to 
the seeds of apricots." The fact that almond-meat is styled "apricot- 
kernel" does not prove that there is a confusion between kin and kin- 
Sen, or between almond and apricot. The confusion may be on the 
part of foreigners who take apricot-kernels for almonds. 2 

It has been stated by Bretschneider 3 that the word pa-lan ffi R 
(*pa-lam), used by the travellers Ye-lu C'u-ts'ai and C'ah C'un, might 
transcribe the Persian word bdddm. This form first appears in the Sun 
H (Ch. 490) in the account of Fu-lin, where the first element is written 
phonetically EL, 4 so that the conclusion is almost warranted that this 
word was transmitted from a language spoken in Fu-lin. In all prob- 
ability, the question is of a Fu-lin word of the type palam or param (per- 
haps *faram, fram, or even *spram). 

The fruit pa-lan must have been known in China during the Sung, 
for it is mentioned by Fan C'en-ta f£ $ j\ (1126-93), in his Kwei hai 
yii hen &', 5 in the description of the §i li Ti J$k (Aleurites triloba), which 

Bretschneider, Early Researches into^the Flora of China, p. 149; Forbes 
and Hemsley, Journal Linnean Soc, Vol. XXIII, p. 217. W. C. Blasdale (Descrip- 
tion of Some Chinese Vegetable Food Materials, p. 48, Washington, 1899) men- 
tions a peculiar variety of the almond imported from China into San Francisco. 
The almond is cultivated in China according to K. v. Scherzer (Berichte osterr. 
Exped. nach Siam, China und Japan, p. 96). L. de Reinach (Le Laos, p. 280) 
states that almond-trees grow in the northern part of Laos. 

2 F. N. Meyer (Agricultural Explorations in the Orchards of China, p. 53) 
supposes erroneously that the consumption of apricot-kernels has given rise to the 
statement that almonds grow in China. Cf. Schlegel's Nederlandsch-Chineesch 
Woordenboek, Vol. I, p. 226. 

3 Mediaeval Researches, Vol. I, p. 20. 

4 Cf. Hirth, China and the Roman Orient, p. 63. His identification with 
Greek &6.\avos y which refers only to the acorn, a wild fruit, is hardly satisfactory, 
for phonetic and historical reasons. For Hirth's translation of ff by "almonds" 
in the same clause read "apricots." 

8 Ed. of Ci pu tsu lai ts'un Su, p. 24. 

The Almond 409 

is said to be like pa-lan-tse. In the Gazetteer of C'en-te fu, pa-lan 'ien 
tl is given as a variety of apricot. 1 

Ho Yi-hin, in his Cen su wen, published in 1884, 2 observes that "at 
present the people of the capital style the almond pa-ta El *i, which is 
identical with pa-tan EL JL. The people of Eastern Ts'i M ^ (San-tuh) 
call the almond, if it is sweet and fine, len hin ^ -& (hazel-nut apricot), 
because it has the taste of hazel-nuts. 3 According to the Hiah tsu pi ki 
^t 18. if tfl, a certain kind of almond, styled 'almond of the I wu hut 
Park' H #7 M $6, is exported from Herat "n & At present it occurs 
in the northern part of China. The fruit offered in the capital is large 
and sweet, that of San-tun is small with thin and scant meat." 

The old tradition concerning the origin of the almond in Persia 
is still alive in modern Chinese authors. The Gazetteer of San-se cou 
in the prefecture of T'ai-p'in, Kwan-si Province, states that the 
flat peach is a cultivation of the country Po-se (Persia). 4 The tree 
is (or was) cultivated in that region. Also the Hwa mu siao ci jfc ^C 
/h 1& (p. 29 b) 5 testifies to indigenous cultivation by saying that almond- 
trees grow near the east side of mountains. It may be, of course, that 
the almond has shared the fate of the date-palm, and that its cultiva- 
tion is now extinct in China. 6 

1 O. Franke, Beschreibung des Jehol-Gebietes, p. 75. 

2 Ch. 12, p. 5 b (see above, p. 399). 

■ This observation is also made by Li Si-£en. 

* San-se lou U _h Jj§. ')H jS, Ch. 14, p. 7 b (published in 1835). 

5 Published in the C'un ts'ao fan tsi ^ ^ ^ ft during the period Tao-kwan 

6 Hauer (Erzeugnisse der Provinz Chili, Mitt. Sent. or. Spr., 1908, p. 14) men- 
tions almonds, large and of sweet flavor, as a product of the district of Mi-yun in Ci-li, 
and both sweet and bitter almonds as cultivated in the district of Lwan-p'in in 
the prefecture of C'en-te (Jehol), the annual output of the latter locality being 
given as a hundred thousand catties, — a hardly credible figure should almonds 
really be involved. Hauer's article is based on the official reports submitted by the 
districts to the Governor-General of the Province in 1904; and the term rendered 
by him "almond" in the original is ta pien fen ^C Jli ^". apparently a local or 
colloquial expression which I am unable to trace in any dictionary. It is at any 
rate questionable whether it has the meaning "almond." O. Franke, in his description 
of the Jehol territory, carefully deals with the flora and products of that region 
without mentioning almonds, nor are they referred to in the Chinese Gazetteer 
of C'eh-te fu. 


42. The fig (Ficus carted) is at present cultivated in the Yang-tse 
valley as a small, irregular shrub, bearing a fruit much smaller and 
inferior in quality to the Persian species. 1 According to the Pen ts'ao 
kan mu, its habitat is Yah-cou (the lower Yang-tse region) and Yun- 
nan. In his time, Li Si-£en continues, it was cultivated also in Ce- 
kiah, Kian-su, Hu-pei, Hu-nan, Fu-kien, and Kwah-tuh (^ ^ IMJ jit) 
by means of twigs planted in the ground. The latter point is of par- 
ticular interest in showing that the process of caprification has remained 
unknown to the Chinese, and, in fact, is not mentioned in their works. 
The fig is not indigenous to China; but, while there is no information in 
Chinese records as to the when and how of the introduction, it is per- 
fectly clear that the plant was introduced from Persia and India, not 
earlier than the T'ang period. 

The following names for the fig are handed down to us: — 

(1) Po-se (Persian) P^I Jf B a-$i, *a-zit(zir) (or H $? a-yi, *a-yik), 2 
corresponds to an Iranian form without n, as still occurs in Kurd heZir 
or ezir. There is another reading, J§H tsan, which is not at the outset 
to be rejected, as has been done by Watters 3 and Hirth. 4 The Pen 
ts'ao kan mu 5 comments that the pronunciation of this character (and 
this is apparently an ancient gloss) should be s& £'w, *dzu, *tsu, *ts'u, 
so that we obtain *adzu, *atsu, *ats'u. This would correspond to an 
ancient Iranian form *aju„ At any rate, the Chinese transcriptions, in 
whatever form we may adopt them, have nothing to do with New 
Persian anjlr, as asserted by Hirth, rjut belong to an older stage of 
Iranian speech, the Middle Persian. 

(2) $k B yin-Zi, 6 *ah-zit(r). This is not "apparently a tran- 

1 Stuart, Chinese Materia Medica, p. 174. The Ci wu mih Si t'u k'ao (Ch. 36, 
p. 2), however, speaks of the fig of Yun-nan as a large tree. According to F. N. 
Meyer (Agricultural Explorations in the Orchards of China, p. 47), the fig is grown 
in northern China only as an exotic, mostly in pots and tubs. In the milder parts of 
the country large specimens are found here and there in the open. He noticed black 
and white varieties. They are cultivated in San-hwa ^ ^ in the prefecture of 
£'an-sa, Hu-nan (San hwa hien U, Ch. 16, p. 15 b, ed. 1877), also in the prefecture 
of Sun-t'ien, Ci-li (Kwan-su Sun t'ien fu li, Ch. 50, p. 10). 

2 Yu yan tsa tsu, Ch. 18, p. 13. 

8 Essays on the Chinese Language, p. 349. 
4 Journal Am. Or. Soc, Vol. XXX, p. 20. 
B Ch. 31, p. 9. 
6 Pen ts'ao kan mu, Ch. 31, p. 26. 


The Fig 411 

scription of Hindustani afijir," as affirmed by Hirth, but of New Persian 
anjtr or enjlr, the Hindustani (as well as Sanskrit anjlra) being simply 
borrowed from the Persian; Bukhara injir, Afghan intsir; Russian 

(3) Fu-lin Ml %M ti-ni or ti-cen & or *® (*ti-tsen, *ti-ten) ; the latter 
variant is not necessarily to be rejected, as is done by Hirth. Cf. 
Assyrian tittu (from *tintu); Phoenician tin; Hebrew ti'nu, te'enah; 1 
Arabic tin, tine, tima; Aramaic ts'intd, tenia, tena; Pahlavi tin (Semitic 
loan-word). The Semitic name is said to have taken its starting-point 
from south-eastern Arabia, where also, in the view of the botanists, the 
origin of fig-culture should be sought; but in view of the Assyrian 
word and the antiquity of the fig in Assyria, 2 this theory is not probable. 
There is no doubt that the Chinese transcription answers to a Semitic 
name; but that this is the Aramaic name, as insisted on by Hirth in 
favor of his theory that the language of Fu-lin should have been Aramaic, 
is not cogent. The transcription ti-ni, on the contrary, is much nearer 
to the Arabic, Phoenician, and Hebrew forms. 3 

(4) 'iE It §fc (or better $£) yu-Van-po, *u-dan-pat(par), *u-dan- 
bar = Sanskrit udambara {Ficus glomerata)* According to Li Si-oen, 
this name is current in Kwah-tun. 

(5) M ^6 :Sv wu hwa kwo ("flowerless fruit"), 5 Japanese icijiku. 
The erroneous notion that the fig-tree does not bloom is not peculiar 
to Albertus Magnus, as Hirth is inclined to think, but goes back to 
times of antiquity, and occurs in Aristotle and Pliny. 6 This wrong 
observation arose from the fact that the flowers, unlike those of most 
fruit-trees, make no outward appearance, but ar,e concealed within the 

1 In the so-called histories of the fig concocted by botanists for popular consump- 
tion, one can still read the absurdity that Latin ficus is to be derived from Hebrew 
feg. Such a Hebrew word does not exist. What does exist in Hebrew, is the word pag, 
occurring only in Canticle (n, 13), which, however, is not a general term for the fig, 
but denotes only a green fig that did not mature and that remained on the tree during 
the winter. Phonetically it is impossible to connect this Hebrew word with the Latin 
one. In regard to the fig among the Semites, see, above all, the excellent article of 
E. Levesque in the Dictionnaire de la Bible (Vol. II, col. 2237). 

2 E. Bonavia, Flora of the Assyrian Monuments, p. 14. 

3 It is surprising to read Hirth's conclusion that "ti-ni is certainly much nearer 
the Aramean word than the Greek owij [better ovkov] for fig, or tpiveds for capri- 
ficus." No one has ever asserted, or could assert, that these Greek words are derived 
from Semitic; their origin is still doubtful (see Schrader in Hehn, Kulturpflanzen, 
p. 100). 

4 Fan yi min yi tsi, Ch. 8, p. 5. 

8 Also other fruits are described under this name (see Ci wu min Si t'u k'ao, 
Ch. 16, pp. 58-60). The terms under 4 and 5 are identified by Kao Si-ki flU ifc "^ 
in his T'ien lu Si yuji j$k Wl f^ (Ch. A, p. 60, published in 1690, ed. of Swo lin). 

6 xvi, 39. 

412 Sino-Iranica 

fruit on its internal surface. On cutting open a fig when it has attained 
little more than one-third its size, the flowers will be seen in full develop- 
ment. 1 

The common fig-tree (Ficus carica) is no less diffused over the Iran- 
ian plateau than the pomegranate. The variety rupestris is found in 
the mountains Kuh-Kiluyeh; and another species, Ficus johannis, 
occurs in Afghanistan between Tebbes and Herat, as well as in Baluchis- 
tan. 2 In the mountain districts of the Taurus, Armenia, and in the 
Iranian table-lands, fig-culture long ago reached a high development. 
Toward the east it has spread to Khorasan, Herat, Afghanistan, as well 
as to Merw and Khiwa. 3 There can be no doubt, either, that the fig was 
cultivated in Sasanian Persia; for it is mentioned in Pahlavi literature 
(above, p. 192), and we have a formal testimony to this effect in the 
Annals of the Liang dynasty, which ascribe udambara to Po-se (Persia) 
and describe the blossoms as charming. 4 In India, as stated, this term 
refers to Ficus glomerata; in China, however, it appears to be also used 
for Ficus carica. Hiian Tsan 5 enumerates udambara among the fruits 
of India. 

Strabo 6 states that in Hyrcania (in Bactria) each fig-tree annually 
produced sixty medimni (one bushel and a half) of fruit. According to 
Herodotus, 7 Croesus was dissuaded from his expedition against Cyrus 
on the plea that the Persians did not even drink wine, but merely water, 
nor did they have figs for sustenance. This, of course, is an anecdote 
without historical value, for we know surely enough that the ancient 
Persians possessed both grapes and wine. Another political anecdote 
of the Greeks is that of Xerxes, who, by having Attic figs served at his 
meals, was daily reminded of the fact that the land where they grow was 
not yet his own. The new discovery of the presence of figs in ancient 
Babylonia warrants the conclusion that they were likewise known and 
consumed in ancient Persia. 

We have no means of ascertaining as to when and how the fig 
spread from Iran to China. The Yu yah tsa tsu is reticent as to the 
transmission, and merely describes the tree as existing in Fu-lin and 

1 Lindley and Moore, Treasury of Botany, pt. 1, p. 492. 

2 C. Joret, Plantes dans l'antiquite\ Vol. II, p. 45. 

8 G. Eisen, The Fig: Its History, Culture, and Curing, p. 20 (U. S. Department 
of Agriculture, Washington, 1901). 

4 Lian Su, Ch. 54, p. 14 b. Read yu-Van-po instead of yu-po-Van, as there printed 
through an oversight. 

6 Ta T"a« si yii ki, Ch. 2, p. 8. 

• II. I, 14. 

7 1, 71. 

The Fig 413 

Persia. 1 We have, however, the testimony of the Arabic merchant Solei- 
man, who wrote in a.d. 851, to the effect that the fig then belonged to 
the fruits of China. 2 

Bretschneider has never written on the subject, but did communicate 
some notes to the botanist Solms-Laubach, from whom they were taken 
over by G. Eisen. 3 Here we are treated to the monstrous statement, 
"The fig is supposed to have reached China during the reign of the 
Emperor Tschang-Kien [sic!], who fitted out an expedition to Turan 
in the year 127 a.d." [sic!]. It is safe to say that Bretschneider could 
not have perpetrated all this nonsense; but, discounting the obvious 
errors, there remains the sad fact that again he credited Can K'ien with 
an introduction which is not even ascribed to him by any Chinese text. 
It is not necessary to be more Chinese than the Chinese, and this 
Changkienomania is surely disconcerting. What a Hercules this Can 
K'ien must have been ! It has never happened in the history of the world 
that any individual ever introduced into any country such a stupendous 
number of plants as is palmed off on him by his epigone admirers. 

Li Si-cen, in his notice of the "flowerless fruit," does not fall back 
on any previous Pen ts'ao; of older works he invokes only the Yu yah 
tsa tsu and the Fan yii U J] % 1&, which mention the udambara of 

The fig of Yun-nan deserves special mention. Wu K'i-tsun, 
author of the excellent botanical work Ci wu mih H Vu k'ao, has de- 
voted a special chapter (Ch. 36) to the plants of Yun-nan, the first of 
these being the yu-Van (udambara) flower, accompanied by two illus- 
trations. From the texts assembled by him it becomes clear that this 
tree was introduced into Yun-nan from India by Buddhist monks. 
Among other stories, he repeats that regarding the monk P'u-t'i(Bodhi)- 
pa-po, which has been translated by C. Sainson; 4 but whereas Yah Sen, 
in his Nan lao ye H f written in 1550, said that one of these trees planted 
by the monk was still preserved in the Temple of the Guardian Spirit 
dfc. i i$! of Yun-nan fu, Wu K'i-tsun states after the Yiin-nan Vuh U 
that for a long time none remained in existence, owing to the ravages 
and burnings of troops. Judging from the illustration, the fig-tree of 
Yun-nan is a species different from Ficus carica. The genus Ficus 

1 Contrary to what is stated by A. de Candolle (Origin of Cultivated Plants, 
p. 296) after Bretschneider. But the description of the fig in that Chinese work 
leaves no doubt that the author speaks from observation, and that the fig, 
accordingly, was cultivated in the China of his time. 

2 M. Reinaud, Relation des voyages, Vol. I, p. 22. 

3 Op. cit., p. 20. 

4 Histoire du Nan-Tchao, p. 196. 

414 Sino-Iranica 

comprises nearly a hundred and sixty species, and of the cultivated fig 
there is a vast number of varieties. 

According to the Yamato-honzd 1 of 1709, figs (iUjiku) were first 
introduced into Nagasaki in the period Kwan-ei % 3K (1624-44) from 
the islands in the South-Western Ocean. This agrees with E. Kaem- 
pfer's 2 statement that figs were brought into Japan and planted by 

1 Ch. 10, p. 26 b. 

2 History of Japan, Vol. I, p. 180 (ed. reprinted Glasgow, 1906). 


43. The Yu yah tsa tsu 1 has the following notice of an exotic plant: 
"The ts'i-t'un ^ ^ (*dzi-tun, *zi-tun) tree has its habitat in the coun- 
try Po-se (Persia), likewise in the country Fu-lin (Syria). In Fu-lin it 
is termed ^ M. ts'i-t'i 2 (*dzi, zi-ti). The tree grows to a height of twenty 
or thirty feet. The bark is green, the flowers are white, resembling 
those of the shaddock (yu $J, Citrus grandis), and very fragrant. 
The fruit is similar to that of the yah-t'ao fa %ft (Averrhoa carambola) 
and ripens in the fifth month. The people of the Western countries 
press an oil out of it for frying cakes and fruit, in the same man- 
ner as sesame seeds {kil-hh E 0) 3 are utilized in China." 

The transcription ts'i-t'un has been successfully identified by Hirth 4 
with Persian zeitun, save that we have to define this form as Middle 
Persian; and Fu-lin tsH-Vi with Aramaic zaitd (Hebrew zayid). This 
is the olive-tree (Olea Europaea). 5 The Persian word is a loan from 
the Semitic, the common Semitic form being *zeitu (Arabic zeitun) . It 
is noteworthy that the Fu-lin form agrees more closely with Grusinian 
and Ossetic zeVi, Armenian jet, dzet ("olive-oil"), zeit ("olive"), Arabic 
zait, 6 than with the Aramaic word. The olive-tree, mentioned in 
Pahlavi literature (above, p. 193), grows spontaneously in Persia and 
Baluchistan, but the cultivated species was in all likelihood received 
by the Iranians (as well as by the Armenians) from the Semites. The 
olive-tree was known in Mesopotamia at an early date: objects in 
clay in the form of an olive belonging to the time of Urukagina, one 
of the pre-Sargonic rulers of Lagash, are still extant. 7 

!Ch. i8,p. 11. 

2 A gloss thus indicates the reading of this character by the fan tsHe HI %•. 

3 See above, p. 292. 

1 Journal Am. Or. Soc, Vol. XXX, 1910, p. 19. 

fi See, for instance, the illustrated article "olivier" in Dujardin-Beaumetz 
and Egasse, Plantes m^dicinales indigenes et exotiques (p. 492, Paris, 1889), which 
is a very convenient and commendable reference-book, particularly valuable for 
its excellent illustrations. Cf. also S. Krauss, Talmudische Archaologie, Vol. II, 
p. 214; S. Fraenkel, Die aramaischen Fremdworter im Arabischen, p. 147. 

8 W. Miller, Sprache der Osseten, p. 10; Hubschmann, Arm. Gram., p. 309. 

7 Handcock, Mesopotamian Archaeology, p. 13. The contributions which 
A. Engler has made to the olive in Hehn's Kulturpflanzen (p. 118) are just as sing- 
ular as his notions of the walnut. Leaves of the olive-tree have been found in Pliocene 
deposits near Mongardino north-west of Bologna, and this is sufficient for Engler 
to "prove" the autochthonous character of the tree in Italy. All it proves, if the 


416 Sino-Iranica 

Schlimmer 1 says that Olea europaea is largely cultivated by the 
inhabitants of Mendjil between Besht and Ghezwin in Persia, and 
that the olives are excellent; nevertheless the oil extracted is very bad 
and unfit to eat. The geographical distribution of the tree in Iran 
has well been traced by F. Spiegel. 2 

The word tsH-Vun has been perpetuated by the lexicographers of 
the Emperor K'ien-lun (1736-95). It makes its appearance in the 
Dictionary of Four Languages, in the section "foreign fruit." 3 For 
the Tibetan and Mongol forms, one has chosen the transcriptions 
Pi-tun siu (transcribing tse -?*) and Utun jimin respectively; while it is 
surprising to find a Manchu equivalent ulusun, which has been correctly 
explained by H. C. v. d. Gabelentz and Sakharov. In the Manchu- 
Chinese Dictionary TsHn wen pu hui, published in 1771, we find the 

fact be correct, is that a wild olive once occurred in the Pliocene of Italy, which 
certainly does not exclude the idea and the well-established historical fact that the 
cultivated olive was introduced into Italy from Greece in historical times. The 
notice of Pliny (xv, 1) weighs considerably more in this case than any alleged 
palasontological wisdom, and the Pliocene has nothing to do with historical times 
of human history. The following is truly characteristic of Engler's uncritical stand- 
point and his inability to think historically: "Since the fruits of the olive-tree are 
propagated by birds, and in many localities throughout the Mediterranean the con- 
ditions for the existence of the tree were prepared, it was quite natural also that the 
tree settled in the localities suitable for it, before the Oriental civilized nations 
made one of the most important useful plants of it." If the birds were the sole 
propagators of the tree, why did they not carry it to India, the Archipelago, and 
China, where it never occurred? The distribution of the olive shows most clearly 
that it was brought about by human activity, and that we are confronted with a 
well-defined geographical zone as the product of human civilization, — Western 
Asia and the Mediterranean area. There is nothing in Engler like the vision and 
breadth of thought of a de Candolle, in whose Origin of Cultivated Plants we read 
(p. 280), "The question is not clearly stated when we ask if such and such olive- 
trees of a given locality are really wild. In ^ woody species which lives so long and 
shoots again from the same stock when cut off by accident, it is impossible to know 
the origin of the individuals observed. They may have been sown by man or birds 
at a very early epoch, for olive-trees of more than a thousand years old are known. 
The effect of such sowing is a naturalization, which is equivalent to an extension 
of area. The point in question is, therefore, to discover what was the home of the 
species in very early prehistoric times, and how this area has grown larger by dif- 
ferent modes of transport. It is not by the study of living olive-trees that this can 
be answered. We must seek in what countries the cultivation began, and how it 
was propagated. The more ancient it is in any region, the more probable it is that 
the species has existed wild there from the time of those geological events which took 
place before the coming of prehistoric man." Here we meet a thinker of critical 
acumen, possessed of a fine historical spirit, and striving for truth nobly and honestly; 
and there, a dry pedant, who thinks merely in terms of species and genera, and is 
unwilling to learn and to understand history. 

1 Terminologie, p. 406. 

2 Eranische Altertumskunde, Vol. I, pp. 257-258. 

3 Appendix, Ch. 3, p. 10. 

The Olive 417 

following definition of ulusun in Chinese: "TsH-Vun is a foreign fruit, 
which is produced in the country Po-se (Persia). The bark of the tree 
is green, the flowers are white and aromatic. Its fruit ripens in the fifth 
month and yields an oil good for frying cakes." This is apparently based 
on the notice of the Yu yah tsa tsu. The Manchu word ulusun (sun 
being a Manchu ending) seems to be an artificial formation based on 
Latin oleum (from Greek elaion), which was probably conveyed through 
the Jesuit missionaries. 

The olive remained unknown to the Japanese; their modern bo- 
tanical science calls it oreiju M W ^ , which reproduces our "olive." 1 
The Japanese botanists, without being aware of the meaning of tsH-tun, 
avail themselves of the characters for this word (reading them ego-no-ki) 
for the designation of Styrax japonica. 2 

The so-called Chinese olive, kan-lan ffli $f, has no affinity with the 
true olive of the West-Asiatic and Mediterranean zone, although its 
appearance comes very near to this fruit. 3 The name kan-lan applies 
to Canarium album and C. pimela, belonging to the order Burseraceae, 
while the olive ranks in that of the Oleaceae* Ma Ci, who, in his K'ai 

1 Matsumura, No. 2136. 

2 Ibid., No. 3051. 

* The kan-lan tree itself is suspected to be of foreign origin; it was most probably 
introduced from Indo-China into southern China. Following are briefly the reasons 
which prompt me to this opinion. 1. According to Li Si-cen, the meaning of the 
name kan-lan remains unexplained, and this comment usually hints at a foreign word. 
The ancient pronunciation was *kam-lam or *kam-ram, which we still find in 
Annamese as kam-lan. The tree abounds in Annam, the fruit being eatable and 
preserved in the same manner as olives (Perrot and Hurrier, Mat. m6d. et phar- 
macop^e sino-annamites, p. 141). Moreover, we meet in Pa-yi, a T'ai language 
spoken in Yun-nan, a word (mak)-k'am, which in a Pa-yi-Chinese glossary is rendered 
by Chinese kan-lan (the element mak means "fruit"; see F. W. K. Miller, T'oung 
Pao, Vol. Ill, p. 27). The relationship of Annamese to the T'ai languages has been 
clearly demonstrated by H. Maspero, and it seems to me that Chinese *kam-lam 
is borrowed from Annam-T'ai. There are many more such Chinese botanical names, 
as I hope to show in the near future. . 2. The plant appears in Chinese records 
at a comparatively recent date. It is first described in the Nan cou i wu li of the 
third century as a plant of Kwah-tuh and Fu-kien and in the Nan fan ts'ao mu Iwan 
(Ch. c, p. 3 b). It is mentioned as a tree of the south in the Kin lou tse of the Em- 
peror Yuan of the Liang in the sixth century (see above, p. 222). A description of 
it is due to Liu Sun in his Lin piao lu i (Ch. B, p. 5 b). In the materia medica it 
first appears in the K'ai pao pen ts'ao of the end of the tenth century. 3. The tree 
remained always restricted to the south-eastern parts of China bordering on Indo- 
China. According to the San fu hwafi t'u, it belonged to the southern plants brought 
to the Fu-li Palace of the Han Emperor Wu after the conquest of Nan Yue (cf. 
above, p. 262). 

* The fruit of Canarium is a fleshy drupe from three to six cm in length, which 
contains a hard, triangular, sharp-pointed seed. Within this are found one or more 
oily kernels. The flesh of the fresh, yellowish-green fruit, like that of the true olive, 
is somewhat acrid and disagreeable, and requires special treatment before it can 

4i 8 Sino-Iranica 

pao pen is'ao (written between a.d. 968 and 976), describes the kan-lan, 
goes on to say that "there is also another kind, known as Pose kan-lan 
('Persian kan-lan'), growing in Yun cou I 'M, 1 similar to kan-lan in 
color and form, but different in that the kernel is divided into two sec- 
tions; it contains a substance like honey, which is soaked in water and 
eaten." The San se lou li % mentions the plant as a product of Sah-se 
cou in Kwah-si. It would be rather tempting to regard this tree as the 
true olive, as tentatively proposed by Stuart; 3 but I am not ready to 
subscribe to this theory until it is proved by botanists that the olive- 
tree really occurs in Kwah-si. Meanwhile it should be pointed out that 
weighty arguments militate against this supposition. First of all, the 
Pose kan-lan is a wild tree: not a word is said to the effect that it is 
cultivated, still less that it was introduced from Po-se. If it had been 
introduced from Persia, we should most assuredly find it as a culti- 
vation; and if such an introduction had taken place, why should it be 
confined to a few localities of Kwah-si? Li Si-cen does not express an 
opinion on the question; he merely says that the Jan ~i) Ian, another 
variety of Canarium to be found in Kwah-si (unidentified), is a kind 
of Pose kan-lan, which proves distinctly that he regards the latter 
as a wild plant. The T'ang authors are silent as to the introduction of 
the olive; nevertheless, judging from the description in the Yu yan tsa 
tsu, it may be that the fruit was imported from Persia under the T'ang. 
Maybe the Pose kan-lan was so christened on account of a certain 
resemblance of its fruit to the olive; we do not know. There is one 
specific instance on record that the Po-se of Ma Ci applies to the 
Malayan Po-se (below, p. 483) ; this may even be the case here, but the 
connection escapes our knowledge. 

S. Julien 4 asserts that the Chinese author from whom he derives 
his information describes the olive-tree and its fruit, but adds that 
the use of it is much restricted. The Chinese name for the tree is not 
given. Finally, it should be pointed out that Ibn Batuta of the four- 
be made palatable. Its most important constituent is fat, which forms nearly one- 
fourth of the total nutritive material. Cf. W. C. Blasdale, Description of Some 
Chinese Vegetable Food Materials, p. 43, with illustration (U. S. Department of 
Agriculture, Bull. No. 68, 1899). The genus Canarium comprises about eighty 
species in the tropical regions of the Old World, mostly in Asia (Engler, Pflan- 
zenfamilien, Vol. Ill, pt. 4, p. 240). 

1 Name under the T'ang dynasty of the present prefecture Nan-nih in Kwah-si 

2 Ch. 14, p. 7 b (see above, p. 409). 

3 Chinese Materia Medica, p. 89. 

4 Industries de l'empire chinois, p. 120. 

The Olive 419 

teenth century positively denies the occurrence of olives in China. 1 
Of course, this Arabic traveller is not an authority on Chinese affairs: 
many of his data concerning China are out and out absurd. He may 
even not have visited China, as suggested by G. Ferrand; notwith- 
standing, he may be right in this particular point. Likewise the Arch- 
bishop of Soltania, who wrote about 1330, states, "There groweth 
not any oil olive in that country." 2 

1 Yule, Cathay, Vol. IV, p. 118. 
1 Ibid., Vol. Ill, p. 96. 


44. In his Pen ts'ao Si i, written during the first half of the eighth 
century, C'en Ts'an-k'i has this notice regarding an exotic plant: 
"A-lo-p'o M i&fft (*a-lak-bwut) grows in the country Fu-lin (Syria), 
its fruit resembling in shape that of the tsao kia -4b 3*$ (Gleditschia or 
Gymnocladus sinensis), save that it is more rounded and elongated. 
It is sweet of taste and savory." 1 

In the Ceh lei pen ts'ao 2 we read that "a-lo-p'o grows in the country 
Fu-§i 1$ jfi"; that is, Bhoja, Sumatra. Then follows the same descrip- 
tion as given above, after C'en Ts'an-k'i. The name p'o-lo-men tsao 
kia §1 M. P3 -fb 3& is added as a synonyme. Li Si-cen 3 comments that 
P'o-lo-men is here the name of a Si-yii M M ("Western Regions") 
country, and that Po-se is the name of a country of the south-western 
barbarians; that is, the Malayan Po-se. The term p'o-lo-men tsao kia, 
which accordingly would mean "Gleditschia of the P'o-lo-men coun- 
try," he ascribes to C'en Ts'an-k'i, but in his quotation from this 
author it does not occur. The country P'o-lo-men here in question is 
the one mentioned in the Man Su* 

A somewhat fuller description of this foreign tree is contained in 
the Yu yah tsa tsu, b as follows: "The Persian tsao kia (Gleditschia) has 
its habitat in the country Po-se (Persia), where it is termed hu-ye- 
yen-mo & *f ft IK, while in Fu-lin it is styled a-li-k'u-fa M M * f£. 6 
The tree has a height of from thirty to forty feet, and measures from 
four to five feet in circumference. The leaves resemble those of Citrus 
medica (kou yuan ffi} mk) , but are shorter and smaller. During the cold 
season it does not wither. 7 It does not flower, and yet bears fruit. 8 
Its pods are two feet long. In their interior are shells (ko ko M Wi). 
Each of these encloses a single seed of the size of a finger, red of color, 

1 Pen ts'ao kan mu, Ch. 31, p. 9 b, where the name of the plant is wrongly 
written a-p'o-lo. The correct form a-lo-p'o is given in the Cen lei pen ts'ao. 

2 Ch. 12, p. 56 (ed. of 1587). 

3 Pen ts'ao kan mu, Ch. 31, p. 9 b. 

4 See below, p. 468. 

6 Ch. 18, p. 12. Also Li Si-gen has combined this text with the preceding one 
under the heading a-p'o-lo (instead of a-lo-p'o). 

6 The Pen ts'ao kan mu (Ch. 31, p. 9 b), in quoting this text, gives the Po-se 
name as hu-ye-yen and the Fu-lin name only as a-li. 

7 This means, it is an evergreen. 

8 This is due to erroneous observation. 


Cassia Pods and Carob 421 

and extremely hard. The interior [the pulp] is as black as [Chinese] 
ink and as sweet as sugar-plums. It is eatable, and is also employed in 
the pharmacopoeia." 

The tree under consideration has not yet been identified, at least not 
from the sinological point of view. 1 The name a-lo-p'o is Sanskrit; and 
the ancient form *a-lak(rak, rag)-bwut(bud) is a correct and logical 
transcription of Sanskrit aragbadha, aragvadha, aragvadha, or drgvadha, 
the Cassia or Cathartocar pus fistula {Leguminosae), already mentioned 
by the physician Caraka, also styled suvarnaka ("gold-colored") and 
rdjataru ("king's tree"). 2 This tree, called the Indian laburnum, 
purging cassia, or pudding pipe tree from its peculiar pods (French 
canificier), is a native of India, Ceylon, and the Archipelago 3 (hence 
Sumatra and Malayan Po-se of the Chinese), "uncommonly beautiful 
when in flower, few surpassing it in the elegance of its numerous long, 
pendulous racemes of large, bright-yellow flowers, intermixed with the 
young, lively green foliage." 4 The fruit, which is common in most 
bazars of India, is a brownish pod, about sixty cm long and two cm 
thick. It is divided into numerous cells, upwards of forty, each con- 
taining one smooth, oval, shining seed. Hence the Chinese comparison 
with the pod of the Gleditschia, which is quite to the point. These pods 
are known as cassia pods. They are thus described in the " Treasury of 
Botany " : " Cylindrical, black, woody, one to two feet long, not splitting, 
but marked by three long furrows, divided in the interior into a number 
of compartments by means of transverse partitions, which project 
from the placentae. Each compartment of the fruit contains a single 
seed, imbedded in pulp, which is used as a mild laxative." Whether 
the tree is cultivated in Asia I do not know; Garcia da Orta affirms 
that he saw it only in a wild state. 5 The description of the tree and 
fruit in the Yu yah tsa tsu is fairly correct. Cassia fistula is indeed 
from twenty to thirty feet high (in Jamaica even fifty feet). The seed, 
as stated there, is of a reddish-brown color, and the pulp is of a dark 
viscid substance. 

1 Stuart (Chinese Materia Medica, p. 496) lists the name a-p'o-lo (instead of 
a-lo-p'o) among "unidentified drugs." Bretschneider has never noted it. 

1 A large number of Sanskrit synonymes for the tree are enumerated by RSdiger 
and Pott {Zeitschrift f. d. K. d. Morg., Vol. VII, p. 154); several more may be added 
to this list from the Bower Manuscript. 

'Garcia da Orta (Markham, Colloquies, p. 114) adds Malacca and Sofala. 
In Javanese it is tenguli or trehgtdi. 

4 W. Roxburgh, Flora Indica, p. 349. 

6 Likewise F. Pyrard (Vol. II, p. 361, ed. of Hakluyt Society), who states that 
"it grows of itself without being sown or tended." 

422 Sino-Iranica 

When I had established the above identification of the Sanskrit 
name, it was quite natural for me to lay my hands on Matsumura's 
"Shokubutsu mei-i" and to look up Cassia fistula under No. 754: 
it was as surprising as gratifying to find there, "Cassia fistula M ffr W) 
namban-saikacki." This Japanese name means literally the " ' Gleditsckia 
japonica {saikali = Chinese tsao-kia-tse) of the Southern Barbarians" 
(Chinese Nan Fan). The Japanese botanists, accordingly, had suc- 
ceeded in arriving at the same identification through the description 
of the plant; while the philological equation with the Sanskrit term 
escaped them, as evidenced by their adherence to the wrong form 
a-p'o-lo, sanctioned by the Pen ts'ao kan mu. The case is of methodo- 
logical interest in showing how botanical and linguistic research may 
supplement and corroborate each other: the result of the identification 
is thus beyond doubt; the rejection of a-p'o-lo becomes complete, and 
the restitution of a-lo-p'o, as handed down in the Cen lei pen ts'ao, 
ceases to be a mere philological conjecture or emendation, but is raised 
into the certainty of a fact. 

The Arabs know the fruit of this tree under the names xarnub hindi 
("Indian carob") 1 and xiyar lanbar ("cucumber of necklaces," from 
its long strings of golden flowers). 2 Abu'l Abbas, styled en-Nebati 
("the Botanist"), who died at Sevilla in 1239, the teacher of Ibn 
al-Baitar, who preserved extracts from his lost work Rihla ("The 
Voyage"), describes Cassia fistula as very common in Egypt, par- 
ticularly in Alexandria and vicinity, whence the fruit is exported to 
Syria; 3 it commonly occurs in Bassora also, whence it is exported to 
the Levant and Irak. He compares the form of the tree to the walnut 
and the fruit to the carob. The same comparison is made by Isak Ibn 
Amran, who states in Leclerc's translation, "Dans chacun de ces tubes 
est renfermee une pulpe noire, sucre\et laxative. Dans chaque com- 
partiment est un noyau qui a le volume et la forme de la graine de 
caroubier. La partie employee est la pulpe, a Pexclusion du noyau et du 

The Persians received the fruit from the Arabs on the one hand, and 
from north-western India on the other. They adopted the Arabic word 
xiyar-Zanbar* in the form xiyar-Zambar (compare also Armenian xiar- 

1 Leclerc, Traits des simples, Vol. II, p. 17. 

2 Ibid., p. 64. Also qitta hindi ("Indian cucumber"), ibid., Vol. Ill, p. 62. 

3 Garcia da Orta says that it grows in Cairo, where it was also found by 
Pierre Belon. In ancient times, however, the tree did not occur in Egypt: Loret, 
in his Flore pharaonique, is silent about it. It was no doubt brought there by the 
Arabs from India. 

4 Garcia da Orta spells it hiar-xamber. 

Cassia Pods and Carob 423 

famb, Byzantine Greek xt-apoo-pfiep, x«a<ra;u7r&p) ; and it is a Middle- 
Persian variation of this type that is hidden in the "Persian" tran- 
scription of the Yu yah tsa tsu, hu-ye-yen-mo %& 5? m R, anciently 
*xut(xur)-ya-dzem(dzem)-m'wak(bak, bax). The prototype to be 
restored may have been *xaryadz'ambax. There is a New-Persian word 
for the same tree and fruit, bakbar. It is also called kabuli ("coming 
from Kabul"). 

The Fu-lin name of the plant is M M :£■ t£ a-li-k'ii-fa, *a-li(ri)- 
go-va5. I. Loew 1 does not give an Aramaic name for Cassia fistula, 
nor does he indicate this tree, neither am I able to find a name for it in 
the relevant dictionaries. We have to take into consideration that the 
tree is not indigenous to western Asia and Egypt, and that the Arabs 
transplanted it there from India (cf. the Arabic terms given above, 
"Indian carob," and "Indian cucumber"). The Fu-lin term is evi- 
dently an Indian loan-word, for the transcription *a-ri-go-va5 cor- 
responds exactly to Sanskrit argvadha, answering to an hypothetical 
Aramaic form *arigbada or *arigfada. In some editions of the Yu yah 
tsa tsu, the Fu-lin word is written a-li or a-li-ja, *a-ri-va5. These would 
likewise be possible forms, for there is also a Sanskrit variant arevata 
and an Indian vernacular form alt (in Panjabl). 

The above texts of C'en Ts'an-k'i and Twan C'en-si, author of 
the Yu yah tsa tsu, give occasion for some further comments. Pelliot 2 
maintained that the latter author, who lived toward the end of the 
ninth century, frequently derived his information from the former, who 
wrote in the first part of the eighth century; 3 from the fact that C'en 
in many cases indicates the foreign names of exotic plants, Pelliot is 
inclined to infer that Twan has derived from him also his nomenclature 
of plants in the Fu-lin language. This is by no means correct. I have 
carefully read almost all texts preserved under the name of C'en (or 
his work, the Pen ts'ao H i) in the Ceh lei pen ts'ao and Pen ts'ao kah mu, 
and likewise studied all notices of plants by Twan; with the result 
that Twan, with a few exceptions, is independent of C'en. As to Fu-lin 
names, none whatever is recorded by the latter, and the above text is 
the only one in which the country Fu-lin figures, while he gives the 
plant-name solely in its Sanskrit form. In fact, all the foreign names 
noted by C'en come from the Indo-Malayan area. The above case 
shows plainly that Twan's information does not at all depend on C'en's 

1 Aramaeische Pflanzennamen. 

2 T'oung Pao, 1912, p. 454. 

3 The example cited to this effect (Bull, de I'Ecole frangaise, Vol. IV, p. 1130) 
is not very lucky, for in fact the two texts are clearly independent. 

424 Sino-Iranica 

passage: the two texts differ both as to descriptive matter and nomen- 
clature. In regard to the Fu-lin information of Twan, Hirth's opinion 1 
is perfectly correct: it was conveyed by the monk Wan, who had 
hailed directly from Fu-lin. 2 The time when he lived is unknown, but 
most probably he was a contemporary of Twan. The Fu-lin names, 
accordingly, do not go back to the beginning of the eighth century, but 
belong to the latter half of the ninth. 

An interesting point in connection with this subject is that both 
the Iranian and the Malayan Po-se play their r61e with reference to 
the plant and fruit in question. This, as far as I know, is the only in- 
stance of this kind. Fortunately, the situation is perfectly manifest on 
either side. The fact that Twan C'eh-si hints at the Iranian Po-se 
(Persia) is well evidenced by his addition of the Iranian name; while 
the tree itself is not found in Persia, and merely its fruit was imported 
from Syria or India. The Po-se, alluded to in the Cen lei pen ts'ao and 
presumably traceable to C'en Ts'an-k'i, unequivocally represents the 
Malayan Po-se: it is joined to the names of Sumatra and P'o-lo-men; 
and Cassia fistula is said to occur there, and indeed occurs in the Malayan 
zone. Moreover, Li Si-cen has added such an unambiguous definition 
of the location of this Po-se, that there is no room for doubt of its identity. 

45. Reference has been made to the similarity of cassia pods to 
carob pods, and it would not be impossible that the latter were included 
in the "Persian Gleditschia" of the Chinese. 

Ceratonia siliqua, the carob-tree, about thirty feet in height, is 
likewise a genus of the family Leguminosae, a typical Mediterranean 
cultivation. The pods, called carob pods, carob beans, or sometimes 
sugar pods, contain a large quantity of mucilaginous and saccharine 
matter, and are commonly employed in the south of Europe for feeding 
live-stock, and occasionally, in times Of scarcity, as human food. The 
popular names "locust-pods" or "St. John's Bread" rest on the suppo- 
sition that the pods formed the food of St. John in the wilderness 
(Luke, xv, 16); but there is better reason to believe that the locusts 
of St. John were the animals so called, and these are still eaten in the 
Orient. The common Semitic name for the tree and fruit is Assyrian 
xarubu, Aramaic xdrubd, Arabic xarrub and xarnub? New Persian 
xurnub (khurnub) or xarnub, also xarrub (hence Osmanli xartip* Neo- 

1 Journal Am. Or. Soc, Vol. XXX, 1910, p. 18. 

2 Cf. above, p. 359. 

3 Egyptian dzarudZ, garuta, darruga; Coptic garate, are Greek loan-words 
(the tree never existed in Egypt, as already stated by Pliny, xin, 16), from Kepdrto. 

4 Also ketHbujnuzu ("goat's horn"). 

Cassia Pods and Carob 425 

Greek xa-poviriov, Italian carrobo or carrubo, Spanish algarrobo, French 
caroube or carouge), is based on the Semitic name. Lelekl is another 
Persian word for the tree, according to Schlimmer, 1 peculiar to Gilan. 

The Arabs distinguish three varieties of carob, two of which are 
named saidalani and fabuni. 2 There is no doubt that the Arabs who 
were active in transplanting the tree to the* west conveyed it also to 
Persia. A. de Candolle does not mention the occurrence of the carob 
in that country. It is pointed out, however, by the Mohammedan 
writers on Persia. It is mentioned as a cultivation of the province 
Sab Or by MuqaddasI 3 and Yaqut. 4 Abu Mansur discusses the medicinal 
properties of the fruit in his pharmacopoeia; he speaks of a Syrian and 
a Nabathaean xarnub? Schlimmer 6 remarks that the tree is very 
common in the forest of Gilan; the pods serve the cows as food, and are 
made into a sweet and agreeable syrup. No Sanskrit name for the 
tree exists, and the tree itself did not anciently occur in India. 7 

A botanical problem remains to be solved in connection with Cassia 
fistula. DuHalde 8 mentions cassia-trees {Cassia fistula) in the province 
of Yun-nan toward the kingdom of Ava. "They are pretty tall, and 
bear long pods; whence 'tis called by the Chinese, Chang-ko-tse-shu, 
the tree with long fruit ( H IP£ ■? m) ; its pods are longer than those we 
see in Europe, and not composed of two convex shells, like those of 
ordinary pulse, but are so many hollow pipes, divided by partitions 
into cells, which contain a pithy substance, in every respect like the 
cassia in use with us." S. W. Williams 9 has the following: "Cassia 
fistula, t*t ^6 flf hwai hwa tsHn, is the name for the long cylindrical pods 
of the senna tree (Cathartocarpus) , known to the Chinese as Van kwo-tse 
$u, or tree with long fruit. They are collected in Kwah-si for their 
pulp and seeds, which are medicinal. The pulp is reddish and sweet, 
and not so drastic as the American sort; if gathered before the seeds 
are ripe, its taste is somewhat sharp. It is not exported, to any great 

1 Terminologie, p. 120. The pods are also styled tarmiS. 

2 L. Leclerc, Trait6 des simples, Vol. II, p. 16. 

3 P. Schwarz, Iran, p. 32. 

4 Barbier de Meynard, Dictionnaire geographique de la Perse, p. 294. 

5 Achundow, Abu Mansur, p. 59. 

6 Terminologie, p. 119. 

7 The alleged word for the carob, gimbibheda, given in the English-Sanskrit 
Dictionary of A. Borooah, is a modern artificial formation from qimbi or gimba 
("pod"). According to Watt, the tree is now almost naturalized in the Salt Range 
and other parts of the Panjab. 

8 Description of the Empire of China, Vol. I, p. 14 (or French ed., Vol. I, p. 26). 

9 Chinese Commercial Guide, p. 114 (5th ed., 1863). 

426 Sino-Iranica 

extent, west of the Cape." F. P. Smith, 1 with reference to this state- 
ment of Williams, asserts that the drug is unknown in Central China, 
and has not been met with in the pages of the Pen ts'ao. Likewise 
Stuart, 2 on referring to DuHalde and Williams, says, "No other 
authorities are found for this plant occurring in China, and it is not 
mentioned in the Pen ts'ao. The Customs Lists do not mention it; so, 
if exported as Williams claims, it must be by land routes. The subject 
is worthy of investigation." Cassia fistula is not listed in the work of 
Forbes and Hemsley. 

There is no doubt that the trees described by DuHalde and Williams 
exist, but the question remains whether they are correctly identified. 
The name hwai used by Williams would rather point to a Sophora, 
which likewise yields a long pod containing one or five seeds, and his 
description of the pulp as reddish does not fit Cassia fistula. Contrary 
to the opinions of Smith and Stuart, the species of Williams is referred 
to in the Pen ts'ao kah mu? As an appendix to his a-p'o-lo (instead of 
a-lo-p x o), Li Si-cen treats of the seeds of a plant styled lo-wan-tse jH 
HI -J*, quoting the Kwei hai yii hen U by Fan C'eh-ta (1126-93) as 
follows: "Its habitat is in Kwan-si. The pods are several inches long, 
and are like those of th&Jei tsao BE 4b {Gleditschia or Gymnocladus sinen- 
sis) and the tao tou 73 fi[ (Canavallia ensijormis). The color [of the 
pulp] is standard red IE ;Q\ Inside there are two or three seeds, which 
when baked are eatable and of sweet and agreeable flavor." 4 This lo-wah 
is identified with Tamarindus indica; 5 and this, I believe, is also the 
above plant of Williams, which must be dissociated from Cassia fistula; 
for, while Li Si-cen notes the latter as a purely exotic plant, he does not 
state that it occurs in China; as to lo-wan, he merely regards it as a 
kindred affair on account of the peculiar pods: this does not mean, of 
course, that the trees yielding these V^pods are related species. The 
fruit of Tamarindus indica is a large swollen pod from four to six inches 
long, filled with an acid pulp. In India it is largely used as food, being 
a favorite ingredient in curries and chutnies, and for pickling fish. It is 
also employed in making a cooling drink or sherbet. 6 

1 Contributions towards the Materia Medica of China, p. 53. 

2 Chinese Materia Medica, p. 96. 

3 Ch. 31, p. 9 b. 

4 The text is exactly reproduced (see the edition in the Ci pu tsu lai ts'un Su, 
p. 24). 

5 Matsumura, No. 3076 (in Japanese cosen-modama-raboU). 
« Watt, Commercial Products of India, p. 1067. 


46. The Yu yah tsa tsu 1 contains the following notice: "The 
habitat of the nai-k'i t^f )® is in the country Fu-lin (Syria). Its sprouts 
grow to a height of three or four feet. Its root is the size of a duck's 
egg. Its leaves resemble those of the garlic (Allium sativum). From the 
centre of the leaves rises a very long stem surmounted by a six-petaled 
flower of reddish-white color. 2 The heart of this flower is yellow-red, and 
does not form fruit. This plant grows in the winter and withers during 
the summer. It is somewhat similar to shepherd's-purse (tsi H, 
Capsella bursa-pastoris) and wheat. 3 An oil is pressed from the flowers, 
with which they anoint the body as a preventive of colds, and is em- 
ployed by the king of Fu-lin and the nobles in his country." 

Li Si-cen, in his Pen ts'ao kan mu* has placed this extract in his 
notice of swi sien ^K fill (Narcissus tazetta), 5 and after quoting it, adds 
this comment: "Judging from this description of the plant, it is similar 
to Narcissus; it cannot be expected, of course, that the foreign name 
should be identical with our own." 6 He is perfectly correct, for the 
description answers this flower very well, save the comparison with 
Capsella. Dioscorides also compares the leaves of Narcissus to those of 
Allium, and says that the root is rounded like a bulb. 7 

The philological evidence agrees with this explanation; for nai-k'i, 
*nai-gi, apparently answers to Middle Persian *nargi, New Persian 
nargis (Arabic narjis), 8 Aramaic narkim, Armenian narges (Persian 

1 Ch. 18, p. 12 b. 

2 Cf. the description of Theophrastus (Hist, plant., vii, 13): "In the case of 
narcissus it is only the flower-stem which comes up, and it immediately pushes up 
the flower." Also Dioscorides (iv, 158) and Pliny (xxi, 25) have given descriptions 
of the flower. 

1 This sentence is omitted (and justly so) in the text, as reprinted in the Pen 
ts'ao kan mu; for these comparisons are lame. 

4 Ch. 13, p. 16. 

s Also this species is said to have been introduced from abroad (Hwa mu siao h 
~%L yfC /J> j£. P- io - b, in £' un is ' ao L' a n ts *> Ch. 25). 

6 In another passage of his work (Ch. 14, p. 10) he has the same text under 
Ian nai |JLf * (Kcempferia galanga), but here he merely adds that the description 
of the Yu yan tsa tsu is "a little like Ian nai." 

7 Leclerc, Traits des simples, Vol. Ill, p. 368. 

8 According to Hubschmann (Armen. Gram., p. 201), the New-Persian form 
would presuppose a Pahlavi *narkis. In my opinion, Greek vbpKicraos is derived from 
an Iranian language through the medium of an idiom of Asia Minor, not vice versd, 
as believed by Noeldeke (Persische Studien, II, p. 43). 


428 Sino-Iranica 

loan-word), denoting Narcissus tazetta, which is still cultivated in 
Persia and employed in the pharmacopoeia. 1 Oil was obtained from the 
narcissus, which is called vapnioouov in the Greek Papyri. 2 

Hirth 3 has erroneously identified the Chinese name with the nard. 
Aside from the fact that the description of the Yu yan tsa tsu does not 
at all fit this plant, his restoration, from a phonetic viewpoint, remains 
faulty. K'ah-hi does not indicate the reading not for the first character, 
as asserted by Hirth, but gives the readings nai, ni, and yin. The second 
character reads k'i, which is evolved from *gi, but does not repre- 
sent ti, as Hirth is inclined to make out. 4 

For other reasons it is out of the question to see the nard in the 
term nai-k'i; for the nard, a product of India, is well known to the 
Chinese under the term kan sun hian ~W t& -ft . 5 The Chinese did not 
have to go to Fu-lin to become acquainted with a product which reached 
them from India, and which the Syrians themselves received from 
India by way of Persia. 6 Hebrew nerd (Canticle), Greek vdpSos, 7 
Persian nard and nard, are all derived from Sanskrit nalada, which 
already appears in the Atharvaveda. 8 Hirth 's case would also run 
counter to his theory that the language of Fu-lin was Aramaic, for 
the word nard does not occur there. 

1 Schlimmer, Terminologie, p. 390. Narcissus is mentioned among the aromatic 
flowers growing in great abundance in Bi§avur, province of Fars, Persia (G. Le 
Strange, Description of the Province of Fars, p. 51). It is a flower much praised 
by the poets Hafiz and Jaml. 

2 T. Reil, Beitrage zur Kenntnis des Gewerbes im hellenistischen Aegypten, 
p. 146. Regarding narcissus-oil, see Dioscorides, 1, 50; and Leclerc, Traits des 
simples, Vol. II, p. 103. 

* Journal Am. Or. Soc, Vol. XXX, 1910, p. 22. 

* See particularly Pelliot, Bull, de I'Ecole frangaise, Vol. IV, p. 291. 
5 Stuart, Chinese Materia Medica, p. 278. 

8 1. Loew, Aram. Pflanzennamen, pp. 368-369. 

7 First in Theophrastus, Hist, plant., IX. vn, 2. 

8 See p. 455. 


47. The Yu yah tsa tsu 1 has the following notice of an exotic plant 
referred exclusively to Syria: "The plant ffl %h ^§ a-p'o-ts'an (*a-bwut- 
sam) has its habitat in the country Fu-lin (Syria). The tree is over ten 
feet high. Its bark is green and white in color. The blossoms are 
fine $H, two being opposite each other (biflorate). The flowers resemble 
those of the rape-turnip, man-tsih Hk W (Brassica rapa-depressa) , 
being uniformly yellow. The seeds resemble those of the pepper-plant, 
hu-tsiao #§ ffl- (Piper nigrum). By chopping the branches, one obtains 
a juice like oil, that is employed as an ointment, serving as a remedy for 
ringworm, and is useful for any disease. This oil is held in very high 
esteem, and its price equals its weight in gold." 

As indicated in the Pen ts'ao kah mu U i, 2 the notice of the plant 
a-p'o-san has been adopted by two works, — the &eh fu t x uh hwi f£ Rb£ 
$£ H", which simply notes that it grows in Fu-lin; and the Hwa i hwa 
mu k"ao 3£ ^ ^£ ^C 3f ("Investigations into the Botany of China and 
Foreign Countries"), which has copied the account of the Yu yah tsa 
tsu without acknowledgment. Neither of these books gives any addi- 
tional information, and the account of the Yu yah tsa tsu remains the 
only one that we possess. 

The transcription *a-bwut(bwur)-sam, which is very exact, leads 
to Aramaic and Talmudic afursama kcditbks (Greek fiaXaanov, 
Arabic balessan), the balm of Gilead (Amyris gileadensis, Balsamoden- 
dron giliadense, or Commiphora opobalsamum, family Burseraceae) of 
ancient fame. This case splendidly corroborates Hirth's opinion that 
the language of Fu-lin (or rather one of the languages of Fu-lin) was 
Aramaic. The last two characters p'o-ts'an (*bwut-sam) could very 
well transcribe Greek balsam; but the element M excludes Greek and 
any other language in which this word is found, and admits no other 
than Aramaic. In Syriac we have apursama and pur soma (pursma), 
hence Armenian aprsam or aprasam* In Neo-Hebrew, qfobalsmon or 

1 Ch. 18, p. 12. 

2 Ch. 4, p. 15. 

3 1. Loew, Aramaeische Pflanzennamen, p. 73. Also afarsma and afarsmon 
(J. Buxtorf, Lexicon chaldaicum, p. 109; J. Levy, Neuhebr. Worterbuch, Vol. I, 
p. 151). Cf. S. Krauss, Talmudische Archaologie, Vol. I, pp. 234-236. 

* HxjBschmann, Armenische Grammatik, p. 107. I do not believe in the Persian 
origin of this word, as tentatively proposed by this author. 


430 Sino-Iranica 

qfofalsmdn is derived from the Greek diro^aXaanov. 1 It is supposed also 
that Old-Testament Hebrew bdsdm refers to the balsam, and might 
represent the prototype of Greek balsamon, while others deny that the 
Hebrew word had this specific meaning. 2 In my opinion, the Greek 
/ cannot be explained from the Hebrew word. 

Twan C'eh-si's description of the tree, made from a long-distance 
report, is tolerably exact. The Amyris gileadensis or balsam-tree is an 
evergreen shrub or tree of the order Amyridaceae, belonging to the 
tropical region, chiefly growing in southern Arabia, especially in the 
neighborhood of Mecca and Medina, and in Abyssinia. As will be seen, 
it was transplanted to Palestine in historical times, and Twan was 
therefore justified in attributing it to Fu-lin. The height of the tree is 
about fourteen feet, with a trunk eight or ten inches in diameter. It 
has a double bark, — an exterior one, thin and red, and an interior one, 
thick and green; when chewed, it has an unctuous taste, and leaves an 
aromatic odor. The blossoms are biflorate, and the fruit is of a gray 
reddish, of the size of a small pea, oblong, and pointed at both ends. 
The tree is very rare and difficult to cultivate. Twan's oil, of course, 
is the light green, fragrant gum exuded from the branches, always highly 
valued as a remedy, especially efficacious in the cure of wounds. 3 It 
was always a very costly remedy, and Twan's valuation (equaling its 
weight in gold) meets its counterpart in the statement of Theophrastus 
that it sells for twice its weight in silver. 

Flavius Josephus (first century a.d.) 4 holds that the introduction 
of the balsam-tree into Palestine, which still flourished there in his 
time, is due to the queen of Saba. In another passage 5 he states that 
the opobalsamum (sap of the tree) grows at Engedi, a city near the lake 
Asphaltitis, three hundred furlongs from Jerusalem; and again, 6 that it 
grows at Jericho: the balsam, he adds in the latter passage, is of all 
ointments the most precious, which, upon any incision made in the wood 
with a sharp stone, exudes out like juice. 

From the time of Solomon it was cultivated in two royal gardens. 

1 J. Levy, op. cit., Vol. I, p. 137. 

2 E. Levesque in Dictionnaire de la Bible, Vol. I, col. 15 17. The rapproche- 
ment of bdsdm and balsamon has already been made by d'Herbelot (Bibliotheque 
orientale, Vol. I, p. 377), though he gives basam only as Persian. The Arabic form 
is derived from the Greek. 

3 Jeremiah, viii, 22. Regarding its employment in the pharmacology of the 
Arabs, see Leclerc, Traite des simples, Vol. I, pp. 255-257. 

4 Antiquitates judaicae, VIII. vi, 6. 
6 Ibid., IX. 1, 2. 

6 Ibid., XIV. iv, 1. 

The Balm of Gilead 431 

This fact was already known to Theophrastus, 1 who gives this account: 
"Balsam grows in the valley of Syria. They say that there are only 
two parks in which it grows, one of about four acres, the other much 
smaller. The tree is as tall as a good-sized pomegranate, and is much 
branched; it has a leaf like that of rue, but it is pale; and it is ever- 
green. The fruit is like that of the terebinth in size, shape, and color, 
and this too is very fragrant, indeed more so than the gum. The gum, 
they say, is collected by making incisions, which is done with bent 
pieces of iron at the time of the Dog-star, when there is scorching heat; 
and the incisions are made both in the trunks and in the upper parts 
of the tree. The collecting goes on throughout the summer; but the 
quantity which flows is not very large: in a day a single man can 
collect a shell-full. The fragrance is exceedingly great and rich, so that 
even a small portion is perceived over a wide distance. However, 
it does not reach us in a pure state: what is collected is mixed with 
other substances; for it mixes freely with such, and what is known in 
Hellas is generally mixed with something else. 2 The boughs are also 
very fragrant. In fact, it is on account of these boughs, they say, that 
the tree is pruned (as well as for a different reason), since the boughs 
cut off can be sold for a good price. In fact, the culture of the trees has 
the same motive as the irrigation (for they are constantly irrigated). 
And the cutting of the boughs seems likewise to be partly the reason 
why the trees do not grow tall; for, since they are often cut about, they 
send out branches instead of putting out all their energy in one direc- 
tion. Balsam is said not to grow wild anywhere. From the larger park 
are obtained twelve vessels containing each about three pints, from the 
other only two such vessels. The pure gum sells for twice its weight 
in silver, the mixed sort at a price proportionate to its purity. Balsam 
then appears to be of exceptional' value." 

As the tree did not occur wild in Palestine, but only in the state of 
cultivation, and as its home is in southern Arabia, the tradition of 
Josephus appears to be well founded, though it is not necessary to 
connect the introduction with the name of the Queen of Saba. 

Strabo, 3 describing the plain of Jericho, speaks of a palace and the 
garden of the balsamum. "The latter," he says, "is a shrub with an 
aromatic odor, resembling the cytisus (Medicago arborea) and the 
terminthus (terebinth-tree) . Incisions are made in the bark, and vessels 

1 Hist, plant., IX, 6 (cf. the edition and translation of A. Hort, Vol. II, p. 245). 

2 E. Wiedemann (Sitzber. phys.-med. Soz. Erl., 1914, pp. 178, 191) has dealt 
with the adulteration of balsam from Arabic sources. 

* XVI. 11, 41. 

432 Sino-Iranica 

are placed beneath to receive the sap, which is like oily milk. When 
collected in vessels, it becomes solid. It is an excellent remedy for head- 
ache, incipient suffusion of the eyes, and dimness of sight. It bears 
therefore a high price, especially as it is produced in no other place." 

Dioscorides 1 asserts erroneously that balsam grows only in a certain 
valley of India and in Egypt; while Ibn al-Baitar, 2 in his Arabic trans- 
lation of Dioscorides, has him correctly say that it grows only' in Judaea, 
in the district called Rur (the valley of the Jordan). It is easily seen 
how Judaea in Greek writing could be misread for India. 

To Pliny, 3 balsamum was only known as a product of Judaea (uni 
terrarum Iudaeae concessum). He speaks of the two gardens after 
Theophrastus, and gives a lengthy description of three different kinds 
of balsamum. 

In describing Palestine, Tacitus 4 says that in all its productions it 
equals Italy, besides possessing the palm and the balsam; and the 
far-famed tree excited the cupidity of successive invaders. Pompey 
exhibited it in the streets of Rome in 65 B.C., and one of the wonderful 
trees accompanied the triumph of Vespasian in a.d. 79. During the 
invasion of Titus, two battles took place at the balsam-groves of Jericho, 
the last being intended to prevent the Jews from destroying the trees. 
They were then made public property, and were placed under the 
protection of an imperial guard; but it is not recorded how long the two 
plantations survived. Tn this respect, the Chinese report of the Yu yah 
tsa tsu is of some importance, for it is apt to teach that the balm of 
Gilead must still have been in existence in the latter part of the ninth 
century. It further presents clear-cut evidence of the fact that 
Judaea was included in the Chinese notion of the country Fu-lin. 

Abd al-Latif (1161-1231) 5 relates how in his time balsam was col- 
lected in Egypt. The operation was preferably conducted in the summer. 
The tree was shorn of its leaves, and incisions were made in the trunk, 
precaution being taken against injuring the wood. The sap was col- 
lected in jars dug in the ground during the heat, then they were taken 
out to be exposed to the sun. The oil floated on the surface and was 
cleaned of foreign particles. This was the true and purest balsam, form- 
ing only the tenth part of the total quantity produced by a tree. At 
present, in Arabia leaves and branches of the tree are boiled. The first 

**, 18. 


2 Leclerc, Traits des simples, Vol. I, 255. 

3 xn, 25, § Hi. 

4 Hist., v, 6. 

5 Silvestre de Sacy, Relation de l'Egypte, p. 20 (Paris, 1810). 

The Balm of Gilead 433 

floating oil is the best, and reserved for the harem; the second is for 

The tree has existed in Egypt from the eleventh to the beginning 
of the seventeenth century. It was presumably introduced there by the 
Arabs. d'Herbelot 1 cites an Arabic author as saying that the balm 
of Mathara near Cairo was much sought by the Christians, owing to 
the faith they put in it. It served them as the chrism in Confirmation. 

The Irish pilgrim Symon Semeonis, who started on his journey to 
the Holy Land in 1323, has the following interesting account of the 
balsam-tree of Egypt: 2 "To the north of the city is a place called 
Matarieh, where is that famous vine said to have been formerly in 
Engaddi (cf. Cant., 1, 13), which distils the balsam. It is diligently 
guarded by thirty men, for it is the source of the greater portion of the 
Sultan's wealth. It is not like other vines, but is a small, low, smooth 
tree, and odoriferous, resembling in smoothness and bark the hazel 
tree, and in leaves a certain plant called nasturtium aquaticum. The 
stalk is thin and short, usually not more than a foot in length; every 
year fresh branches grow out from it, having from two to three feet in 
length and producing no fruit. The keepers of the vineyard hire Chris- 
tians, who with knives or sharp stones break or cut the tops of these 
branches in several places and always in the sign of a cross. The balsam 
soon distils through these fractures into glass bottles. The keepers 
assert that the flow of balsam is more abundant when the incision 
is made by a Christian than by a Saracen." 3 

In 1550 Pierre Belon 1 still noted the tree in Cairo. Two speci- 
mens were still alive ini6i2. Ini6i5, however, the last tree died. 

The Semitic word introduced into China by the Yu yah tsa tsu 
seems to have fallen into oblivion. It is not even mentioned in the 
Pen ts'ao kan mu. The word "balsam," however, was brought back to 
China by the early Jesuits. In the famous work on the geography of 
the world, the Cifah wai ki $$ ~j) 9V $6, 5 first draughted by Pantoja, and 
after his death enlarged and edited in 1623 by Giulio Aleni (1582-1649), 
the Peru balsam is described under the name pa'r-sa-mo $}L M ffli 0. 
The same word with reference to the same substance is employed by 

1 Bibliotheque orientale, Vol. I, p. 392. 

2 M. Esposito, The Pilgrimage of Symon Semeonis: A Contribution to the 
History of Mediaeval Travel (Geographical Journal, Vol. LI, 1918, p. 85). 

3 Cf. the similar account of K. v. Megenberg (Buch der Natur, p. 358, writ- 
ten in 1349-50). 

4 Observations de plusieurs singularitez et choses memorables, trouv^es en 
Grece, Asie, Iud6e, Egypte, Arabie, p. 246. 

5 Ch. 4, p. 3 (ed. of Sou San ko ts'un Im). 

434 Sino-Iranica 

Ferdinand Verbiest (1623-88) in his K'un yii fu $wo fy M M 1£, and 
was hence adopted in the pharmacopoeia of the Chinese, for it figures 
in the Pen ts x ao kan mu H i. 1 The Chinese Gazetteer of Macao 2 mentions 
pa 'r-su-ma aromatic EL W 8£ & ^ as a kind of benjoin. In this case 
we have a transcription of Portuguese bdlsamo. 

1 Ch. 6, p. 19. See, further, Waiters, Essays on the Chinese Language, p. 339. 

2 Ao-men ci Uo, Ch. B, p. 41 (cf. Wylie, Notes on Chinese Literature, p. 60). 


48. The preceding notes on Fu-lin plants have signally confirmed 
Hirth's opinion in regard to the language of Fu-lin, which was Aramaic. 
There now remains but one Fu-lin plant-name to be identified. This is 
likewise contained in the Yu yah tsa tsu. 1 The text runs as follows: — 

"The p'an-nu-se ^*ili§ $§ tree has its habitat in Po-se (Persia), 
likewise in Fu-lin. In Fu-lin it is styled k'un-han ffl ft. The tree is 
thirty feet high, and measures from three to four feet in circumference. 
Its leaves resemble those of the si $uh iftfl $? (the Banyan tree, Ficus 
retusa). It is an evergreen. The flowers resemble those of the citrus, 
kii tSj and are white in color. The seeds are green and as large as a 
sour jujube, swan tsao §§ 31 (Diospyros lotus) . They are sweet of taste 
and glossy (fat, greasy) . They are eatable. The people of the western 
regions press oil out of them, to oint their bodies with to ward off 

The transcription p % an-nu-se answers to ancient *bwan-du-sek; 
and k'iin-han, to ancient g'win-xan. Despite a long-continued and 
intensive search, I cannot discover any Iranian plant-name of the type 
bandusek or wandusek, nor any Aramaic word like ginxan. The botanical 
characteristics are too vague to allow of a safe identification. Never- 
theless I hope that this puzzle also will be solved in the future. 2 

In the Fu-lin name a-li-k'ii-fa we recognized an Indian loan-word in 
Aramaic (p. 423). It would be tempting to regard as such also the 
Fu-lin word for "pepper" *a-li-xa-da H M ffi P£ (a-li-ho-t'o), which 
may be restored to *alixada, arixada, arxad; but no such word is known 
from Indian or in Aramaic. The common word for " pepper " in Aramaic 
is filfol (from Sanskrit pippala) . In certain Kurd dialects J. de Morgan 3 
has traced a word alat for "pepper," but I am not certain that this is 

1 Ch. i8, p. 10 b. 

2 My colleague, Professor M. Sprengling at the University of Chicago, kindly 
sent me the following information: "Olive-oil was used to ward off ulcers (see 
Winer, Bibl. Realwortb., Vol. II, p. 170; and Krauss, Archaeologie des Talmud, 
Vol. I, pp. 229, 233, 683). Neither in Krauss nor elsewhere was I able to find the 
name of an oil-producing tree even remotely resembling ginxan. There is a root 
qnx ('to wipe, to rub, to anoint'). It is theoretically possible that gis pronounced 
voiced and thus becomes a guttural g, and that from this root, by means of the 
suffix -an, may be derived a noun *qlnxan, *ginxan to which almost any significance 
derived from 'rubbing, anointing* might be attached. But for the existence of such 
a noun or adjective I have not the slightest evidence." 

3 Mission scientifique en Perse, Vol. V, p. 132. 


436 Sino-Iranica 

connected with our Fu-lin word, which at any rate represents a loan- 

There is another Fu-lin word which has not yet been treated cor- 
rectly. The T'ang Annals, in the account of Fu-lin (Ch. 221), mention 
a mammal, styled ts'un If, of the size of a dog, fierce, vicious, and 
strong. 1 Bretschneider, 2 giving an incorrect form of the name, has 
correctly identified this beast with the hyena, which, not being found 
in eastern Asia, is unknown to the Chinese. Ma Twan-lin adds that 
some of these animals are reared, 3 and the hyena can indeed be tamed. 
The character for the designation of this animal is not listed in K'ah-hi's 
Dictionary; but K'an-hi gives it in the form U 4 with the pronunciation 
Men (fan-tsHe j£ %&, sound equivalent Hi), quoting a commentary to 
the dictionary Er ya, which is identical with the text of Ma Twan-lin 
relative to the animal ts'un. This word Men (or possibly hilan) can be 
nothing but a transcription of Greek vaiva, hyaena, or vaivrj. On the 
other hand, it should be noted that this Greek word has also passed as 
a loan into Syriac; 5 and it would therefore not be impossible that it 
was Syrians who transmitted the Greek name to the Chinese. This 
question is altogether irrelevant; for we know, and again thanks to 
Hirth's researches, that the Chinese distinguished two Fu-lin, — the 
Lesser Fu-lin, which is identical with Syria, and the Greater Fu-lin, the 
Byzantine Empire with Constantinople as capital. 6 Byzantine Greek, 
accordingly, must be included among the languages spoken in Fu-lin. 

As to the origin of the name Fu-lin, I had occasion to refer to Pel- 
liot's new theory, according to which it would be based on Rom, 
Rum. 7 I am of the same opinion, and perfectly in accord with the 
fundamental principles by which this theory is inspired. In fact, this 
is the method followed throughout this investigation: by falling 
back on the ancient phonology of Chinese, we may hope to restore 
correctly the prototypes of the Chinese transcriptions. Pelliot starts 
from the Old-Armenian form Hrom or HrOm, 8 in which h represents 

1 Hirth, China and the Roman Orient, pp. 60, 107, 220. 

2 Knowledge possessed by the Ancient Chinese of the Arabs, p. 24. 

3 Hirth (op. cit., p. 79) translates, "Some are domesticated like dogs." But 
the phrase f£[ $»J following W ^ ^ forms a separate clause. In the text printed 
by Hirth (p. 115, Q 22) the character ~j] is to be eliminated. 

4 Thus reproduced by Palladius in his Chinese-Russian Dictionary (Vol. I, 
p. 569) with the reading silan. 

6 R. P. Smith, Thesaurus syriacus, Vol. I, col. 338. 

Cf. Hirth, Journal Am. Or. Soc, Vol. XXXIII, 1913, pp. 202-208. 

7 The Diamond (this volume, p. 8). Pelliot's notice is in Journal asiatique, 
1914, I, pp. 498-500. 

8 Cf. HtfBSCHMANN, Armen. Gram., p. 362. 

Note on the Language of Fu-lin 437 

the spiritus asper of the initial Greek r. In some Iranian dialects the 
spiritus asper is marked by an initial vowel: thus in Pahlavi Arum, in 
Kurd Urum. The ancient Armenian words with initial hr, as explained 
by A. Meillet, were borrowed from Parthian dialects which transformed 
initial Iranian / into h: for instance, Old Iranian jramana (now fertnan, 
"order") resulted in Armenian hraman, hence from Parthian *hraman. 
Thus *FrOm, probably conveyed by the Sogdians, was the prototype 
from which Chinese Fu-lin, *Fu-lim, was fashioned. In my opinion, 
the Chinese form is not based on *FrOm, but on *Frim or *Frfm. Rim 
must have been an ancient variant of Rum; Rim is still the Russian 
designation of Rome. 1 What is of still greater importance is that, as 
has been shown by J. J. Modi, 2 there is a Pahlavi name Sairima, which 
occurs in the Farvardin Yast, and is identified with Rum in the Bun- 
dahisn; again, in the Sahnameh the corresponding name is Rum. This 
country is said to have derived its name from Prince Selam, to whom 
it was given; but this traditional opinion is not convincing. A form 
Rima or Rim has accordingly existed in Middle Persian; and, on the 
basis of the Chinese transcription *Fu-lim or *Fu-rim, it is justifiable 
to presuppose the Iranian (perhaps Parthian) prototype *Frim, from 
which the Chinese transcription was made. 

1 What Pelliot remarks on the Tibetan names Ge-sar and P'rom is purely 
hypothetical, and should rather be held in abeyance for the present. We know so 
little about the Ge-sar epic, that no historical conclusions can be derived from it. 
For the rest, the real Tibetan designation for Byzance or Turkey, in the same 
manner as in New Persian, is Rum (T'oung Pao, 1916, p. 491). In regard to the 
occurrence of this name in Chinese transcriptions of more recent date, see Bret- 
schneider, Mediaeval Researches, Vol. II, p. 306; and Hirth, Chau Ju-kua, p. 141. 

2 Asiatic Papers, p. 244 (Bombay, 1905). 


49. This Cucurbitacea (Citrullus vulgaris or Cucurbita citrullus) 
is known to the Chinese under the name si kwa IS id ("melon of the 
west"). The plant now covers a zone from anterior Asia, the Caucasus 
region, Persia to Turkistan and China, also southern Russia and the 
regions of the lower Danube. There is no evidence to lead one to sup- 
pose that the cultivation was very ancient in Iran, India, Central Asia, 
or China; and this harmonizes with the botanical observation that 
the species has not been found wild in Asia. 1 

A. Engler 2 traces the home of the water-melon to South Africa, 
whence he holds it spread to Egypt and the Orient in most ancient times, 
and was diffused over southern Europe and Asia in the pre-Christian 
era. This theory is based on the observation that the water-melon 
grows spontaneously in South Africa, but it is not explained by what 
agencies it was disseminated from there to ancient Egypt. Neverthe- 
less the available historical evidence in Asia seems to me to speak 
in favor of the theory that the fruit is not an Asiatic cultivation; and, 
since there is no reason to credit it to Europe, it may well be traceable 
to an African origin. 

The water-melon is not mentioned by any work of the T'ang dy- 
nasty; notably it is absent from the T'ai pHn hwan yii ki. The earliest 
allusion to it is found in the diary of Hu Kiao jijE HI, entitled Hien lu ki 
fiS M 12, which is inserted in chapter 73 of the History of the Five Dy- 
nasties (Wu tai $i), written by Nou-yan Siu Hfc $0 j£ (a.d. 1017-72) 
and translated by E. Chavannes. 3 Hu Kiao travelled in the country 
of the Kitan from a.d. 947 to 953, and narrates that there for the first 
time he ate water-melons (si kwa). 4 He goes on to say, "It is told that 
the Kitan, after the annihilation of the Uigur, obtained this cultivation. 
They cultivated the plant by covering the seeds with cattle-manure 
and placing mats over the beds. The fruit is as large as that of the 

1 A. de Candolle, Origin of Cultivated Plants, p. 263. 

2 In Hehn, Kulturpfianzen, p. 323. 

* Voyageurs chinois chez les Khitan (Journal asiatique, 1897, I, pp. 390-442). 

4 Chavannes' translation "melons" (p. 400) is inadequate; the water-melon 
is styled in French pasteque or melon d'eau. Hu Kiao, of course, was acquainted 
with melons in general, but what he did not previously know is this particular species. 
During Napoleon's expedition to Egypt, "on mangeait des lentilles, des pigeons, et 
un melon d'eau exquis, connu dans les pays m^ridionaux sous le nom de pasteque. 
Les soldats l'appelaient sainte pasteque' ' (Thiers, Histoire de la revolution f rancaise). 


The Water-Melon 439 

tun kwa %- Jtt. (Benincasa cerifera) 1 and of sweet taste." 2 The water- 
melon is here pointed out as a novelty discovered by a Chinese among 
the Kitan, who then occupied northern China, and who professed to 
have received it from the Turkish tribe of the Uigur. It is not stated 
in this text that Hu Kiao took seeds of the fruit along or introduced it 
into China proper. This should be emphasized, in view of the con- 
clusion of the Pen ts'ao kan mu (see below), and upheld by Bretschneider 
and A. de Candolle, that the water-melon was in China from the tenth 
century. At that time it was only in the portion of China held by the 
Kitan, but still unknown in the China of the Chinese. 3 

1 "Cultivated in China, Japan, India and Africa, and often met with in a wild 
state: but it is uncertain whether it is indigenous" (Forbes and Hemsley, Journal 
Linnean Society, Vol. XXIII, p. 315). 

2 Hu Kiao was a good observer of the flora of the northern regions, and his 
notes have a certain interest for botanical geography. Following his above refer- 
ence to the water-melon, he continues, "Going still farther east, we arrived at Niao- 
t'an, where for the first time willows [Jurci suxei] are encountered, also water-grass, 
luxuriant and fine; the finest of this kind is the grass si-ki Jj, f?| with large blades. 
Ten of these are sufficient to satisfy the appetite of a horse. From Niao-t'an we 
advanced into high mountains which it took us ten days' journey to cross. Then we 
passed a large forest, two or three /*' long, composed entirely of elms, wu-i $jfe ^ 
( Ulmus macrocarpa) , the branches and leaves of which are set with thorns like arrow- 
feathers. The soil is devoid of grass." Si-ki apparently represents the transcription 
of a Kitan word. Three species of elm occur in the Amur region, — Ulmus montana, 
U. campestris, and U. suberosa (Grum-Grzimailo, Opisanie Amurskoi Oblasti, 
p. 316). In regard to the locality T'an-6'eh-tien, Hu Kiao reports, "The climat 
there is very mild, so that the Kitan, when they suffer from great cold, go there to 
warm up. The wells are pure and cool; the grass is soft like down, and makes a 
good sleeping-couch. There are many peculiar flowers to be found, of which two 
species may be mentioned, — one styled han-kin ^ ^, the size of the palm of a 
hand, of gold color so brilliant that it dazzles man; the other, termed ts'in Zan 
"fff j|£ like the kin t'en & $£ (Orithia edulis) of China, resembling in color an 
Indigofera (Ian ^§f) and very pleasing." The term han-kin appears to be the tran- 
scription of a Kitan word; so is perhaps also ts'in £an, although, according to Stuart 
(Chinese Materia Medica, p. 404), the leaves of Sesamum are so called; this plant, 
however, cannot come here into question. 

* The Pien tse lei pien cites the Wu tai Si to th$ effect that Siao Han |§f tjfo, 
after the subjugation of the Uigur, obtained the seeds of water-melons and brought 
them back, and that the fruit as a product of the Western Countries (Si yu, that is, 
Central Asia) was called "western melon" (si kwa). I regret not having been able to 
trace this text in the Wu tai Si. The biography of Siao Han inserted in the Kiu 
Wu tai Si (Ch. 98, pp. 6 b-7 a) contains nothing of the kind. The statement itself 
is suspicious for two reasons. Siao Han, married to A-pu-li, sister of the Emperor 
Wu-yu, in a.d. 948 was involved in a high-treason plot, and condemned to death in 
the ensuing year (cf. H. C. v. D. Gabelentz, Geschichte der grossen Liao, p. 65; 
and Chavannes, op. cit., p. 392). Hu Kiao was secretary to Siao Han, and in this 
capacity accompanied him to the Kitan. After his master's death, Hu Kiao was 
without support, and remained among the Kitan for seven years (up to the year 953). 
It was in the course of these peregrinations that, as related above, he was first 
introduced to water-melons. Now, if Siao Han had really introduced this fruit into 

440 Sino-Iranica 

The man who introduced the fruit into China proper was Hun Hao 
$c fiS (a.d. 1090-1155), ambassador to the Kin or Jurei, among whom he 
remained for fifteen years (1129-43). In his memoirs, entitled Sun mo 
ki wen t& iH $S W, he has the following report: 1 "The water-melon 
(si kwa) is in shape like a flat Acorus (p'u W), but rounded. It is very 
green in color, almost blue-green. In the course of time it will change 
into yellow. This Cucurbitacea (Vie 1&£) resembles the sweet melon (Hen 
kwa ^ft JR, Cucumis melo), and is sweet and crisp. 2 Its interior is filled 

China during his lifetime (that is, prior to the year 949), we might justly assume 
that his secretary Hu Kiao must have possessed knowledge of this fact, and would 
hardly speak of the fruit as a novelty. Further, the alleged introduction of the 
fruit by Siao Han conflicts with the tradition that this importation is due to Huh 
Hao in the twelfth century (see above). It would be nothing striking, of course, if, as 
the fruit was cultivated by the Kitan, several Chinese ambassadors to this people 
should have carried the seeds to their country; but, as a rule, such new acquisitions 
take effect without delay, and if Siao Han had imported the seeds, there was no 
necessity for Huh Hao to do so again. Therefore it seems preferable to think either 
that the text of the above quotation is corrupted, or that the tradition, if it existed, 
is a subsequent makeshift or altogether erroneous. 

1 Not having access to an edition of this work, I avail myself of the extract, as 
printed in the Kwan k'iinfan P'u (Ch. 14, p. 17 b), the texts of which are generally 
given in a reliable form. 

2 In regard to the melon (Cucumis melo), A. de Candolle (Origin of Cultivated 
Plants, p. 261) says with reference to a letter received from Bretschneider in 1881, 
"Its introduction into China appears to date only from the eighth century of our 
era, judging from the epoch of the first work which mentions it. As the relations 
of the Chinese with Bactriana, and the north-west of India by the embassy of 
Chang-Kien, date from the second century, it is possible that the culture of the 
species was not then widely diffused in Asia." Nothing to the effect is to be found in 
Bretschneider's published works. In his Bot. Sin. (pt. II, p. 197) he states that all 
the cucurbitaceous plants now cultivated for food in China are probably indigenous 
to the country, with the exception of the cucumber and water-melon, which, as their 
Chinese names indicate, were introduced from the West. In the texts assembled 
in the Pen ts'ao kan mu regarding tien kwa,\no allusion is made to foreign origin. 
Concerning the gourd or calabash (Lagenariu vulgaris), A. de Candolle (I. c, 
p. 246) states after a letter of Bretschneider that "the earliest work which mentions 
the gourd is that of Tchong-tchi-chou, of the first century before Christ, quoted in 
a work of the fifth or sixth century." This seems to be a confusion with the Cun 
Su $u of the T'ang period (Bretschneider, Bot. Sin., pt. I, p. 79). The gourd, of 
course, occurs in ancient canonical literature (Bot. Sin., pt. II, p. 198). The history 
of this and other cucurbitaceous plants requires new and critical investigation, the 
difficulty of which is unfortunately enhanced by a constant confusion of terms in 
all languages, the name of one species being shifted to another. It means very little, 
of course, that at present, as recently emphasized again by H. J. Spinden (Pro- 
ceedings Nineteenth Congress of Americanists, p. 271, Washington, 1917), Lagenaria 
is distributed over the New and Old Worlds alike; the point is, where the centre of the 
cultivation was (according to A. de Candolle it was in India; see, further, Asa Gray, 
Scientific Papers, Vol. I, p. 330), and how it spread, or whether the wild form had a 
wide geographical range right from the beginning, and was cultivated independently 
in various countries. In view of the great antiquity of the cultivation both in India 
and China, the latter assumption would seem more probable; but all this requires 
renewed and profound investigation. 

The Water-Melon 441 

with a juice which is very cold. Hun Hao, when he went out as envoy, 
brought the fruit back to China. At present it is found both in the 
imperial orchards and in village gardens. It can be kept for several 
months, aside from the fact that there is nothing to prevent it from 
assuming a yellow hue in course of time. In P'o-yah 3K fi§ 1 there lived 
a man who for a long time was afflicted with a disease of the eyes. 
Dried pieces of water-melon were applied to them and caused him relief, 
for the reason that cold is a property of this fruit." Accordingly the 
water-melon was transplanted into China proper only in the latter 
part of the twelfth century. Also the Si wu ki yiian ^ # ^ i^, 2 which 
says that in the beginning there were no water-melons in China, 
attributes their introduction to Hun Hao. The Kin or Jurci, a nation 
of Tungusian origin, appear to have learned the cultivation from the 
Kitan. From a JurSi-Chinese glossary we know also the JurcH designa- 
tion of the water-melon, which ^is xeko, corresponding to Manchu 
xengke, a general term for cucurbitaceous plants. In Golde, xinke 
(in other Tungusian dialects kemke, kenke) denotes the cucumber, and 
seho or sego the water-melon. The proper Manchu word for the water- 
melon is dungga or dunggan. The Tungusian tribes, accordingly, did 
not adopt the Persian-Turkish word karpuz (see below) from the Uigur, 
but applied to the water-melon an indigenous word, that originally 
denoted another cucurbitaceous species. 

Following is the information given on the subject in the Pen ts'ao 
kan mu. 

Wu 2ui zik 2Si, a physician from the province of Ce-kiah in the 
thirteenth century, author of the 7A yun pen ts'ao ffi ^ ^, is cited 
in this work as follows: "When the Kitan had destroyed the Uigur, 
they obtained this cultivation. They planted this melon by covering 
the seeds with cattle-manure. The formation of this fruit is like the 
peck tou ^1*; it is large and round like a gourd, and in color like green 
jade. The seeds have a color like gold, but some like black hemp. In 
the northern part of our country the fruit is plentiful." Li Si-cen ob- 
serves, "According to the Hien lu ki by Hu Kiao (see p. 438), this 
cultivation was obtained after the subjugation of the Uigur. It is styled 
'western melon' (si kwa). Accordingly it is from the time of the Wu-tai 
(a.d. 907-960) that it was first introduced into China. 3 At present it 
occurs both in the south and north of the country, though the southern 

1 In the prefecture of 2ao-cou, Kian-si. 

2 The work of Kao C'eh jti ZJSc of the Sung dynasty. 

3 The same opinion is expressed by Yan Sen (1 488-1 559) in his Tan k'ien tsun 
lu (above, p. 331). 

442 Sino-Iranica 

fruit is inferior in taste to that of the north." He distinguishes sweet, 
insipid, and sour varieties. 

In the T'ao hun kin lu ¥$ ^ 2r i£ l it is stated that in Yun-kia 
3R H (in the prefecture of Wen-cou, Ci-li) there were han kwa ^ jK. 
("cold melons") of very large size, which could be preserved till the 
coming spring, and which are regarded as identical with the water- 
melon. Li Si-Sen justly objects to this interpretation, commenting that, 
if the water-melon was first introduced in the Wu-tai period, the name 
si kwa could not have been known at that time. This objection must 
be upheld, chiefly for the reason that we have no other records from the 
fourth century or even the T'ang period which mention the water- 
melon: it is evidently a post-T'ang introduction. 2 

Ye Tse-k'i, in his Ts'ao mu tse W-JIs-f* written in 1378, remarked 
that water-melons were first introduced under the Yuan, when the 
Emperor Si-tsu tfr JfiH. (Kubilai) subjugated Central Asia. This view 
was already rejected under the Ming in the Cen cu c'wan & *%. $& by 
C'en Ki-zu $1 $1 Hf , who aptly referred to the discovery of the fruit by 
Hu Kiao, and added that it is not mentioned in the Er ya, the various 
older Pen ts'ao, the TsH min yao Su, and other books of a like character, 
it being well known that the fruit did not anciently exist in China. As 
to this point, all Chinese writers on the subject appear to be agreed; and 
its history is so well determined, that it has not given rise to attempts 
of antedating or "changkienizing" the introduction. 

The Chinese travellers during the Mongol period frequently allude 
to the large water-melons of Persia and Central Asia. 3 On the other 
hand, Ibn Batata mentions the excellent water-melons of China, which 
are like those of Khwarezm and Ispahan. 4 

According to the Manchu officers Fusambd and Surde, who pub- 
lished an account of Turkistan about, 1772, 5 the water-melon of this 
region, though identical with that of China, does not equal the latter 
in taste; on the contrary, it is much inferior to it. Other species of melon 
belong to the principal products of Turkistan; some are called by the 
Chinese "Mohammedan caps" and "Mohammedan eyes." The so- 
called "Hami melon," which is not a water-melon, and ten varieties 
of which are distinguished, enjoys a great reputation. Probably it is 

1 Apparently a commentary to the works of T'ao Hun-kin (a.d. 451-536). 

2 The alleged synonyme han kwa for the water-melon, adopted also by Bret- 
schneider {Chinese Recorder, 1871, p. 223) and others, must therefore be weeded out. 

3 Cf. Bretschneider, Mediaeval Researches, Vol. I, pp. 20, 31, 67, 89. 

4 Yule, Cathay, new ed., Vol. IV, p. 109. 

8 Hui k'ian U, see above, p. 230; and below, p. 562. 

The Water-Melon 443 

a variety of sweet melon {Cucumis meld), called in Uigur and Djagatai 
kogun, kavyn, or kaun, in TurkI qawa and qawaq. 

It is said to have been introduced into China as late as the K'an-hi 
era (1662-1721), and was still expensive at that time, but became 
ubiquitous after the subjugation of Turkistan. 1 Of other foreign 
countries that possess the water-melon, the Yin yai Sen Ian mentions 
Su-men-ta-la (Sumatra), where the fruit has a green shell and red 
seeds, and is two or three feet in length, 2 and Ku-li l§f M (Calicut) in 
India, where it may be had throughout the year. 3 In the country of the 
Mo-ho the fruits are so heavy that it takes two men to lift them. They 
are said to occur also in Camboja. 4 If it is correct that the first report 
of the water-melon reached the Chinese not earlier than the tenth 
century (and there is no reason to question the authenticity of this 
account), this late appearance of the fruit would rather go to indicate 
that its arrival in Central Asia was almost as late or certainly not much 
earlier; otherwise the Chinese, during their domineering position in 
Central Asia under the T'ang, would surely not have hesitated to 
appropriate it. This state of affairs is confirmed by conditions in Iran 
and India, where only a mediaeval origin of the fruit can be safely sup- 

The point that the water-melon may have been indigenous in 
Persia from ancient times is debatable. Such Persian terms as hindewane 
("Indian fruit") [Afghan hindwdna] or battix indi ("Indian melon") 5 
raise the suspicion that it might have been introduced from India. 6 
Garcia da Orta states, "According to the Arabs and Persians, this 
fruit was brought to their countries from India, and for that reason they 

1 Hui k'ian li, Ch. 2; and Ci wu min H t'u k'ao, Ch. 16, p. 85. 

2 Malayan mandelikei, tambikei, or semanka (Javanese semonka, Cam samkai). 
Regarding other Malayan names of cucurbitaceous plants, see R. Brandstetter, 
Mata-Hari, p. 27; cf. also J. Crawfurd, History of the Indian Archipelago, Vol. I, 
P- 435- 

s Regarding other cucurbitaceous plants of Calicut, see Rockhill, T'oung Pao, 
1915, pp. 459, 460; but tun kwa is not, as there stated, the cucumber, it is Benincasa 

4 Kwan k'un fan p'u, Ch. 14, p. 18. Cf. Pelliot, Bull, de VEcole franqaise, 
Vol. II, p. 169. Water-melons are cultivated in Siam (Pallegoix, Description 
du royaume Thai, Vol. I, p. 126). 

4 From the Arabic; Egyptian bettu-ka, Coptic betuke; hence Portuguese and 
Spanish pasteca, French pastique. The battix hindi has already been discussed by Ibn 
al-Baitar (L. Leclerc, Traits des simples, Vol. I, p. 240) and by Abu Mansur (Achun- 
dow, p. 23). Armenian ttum bears no relation to the dudaim of the Bible, as tenta- 
tively suggested by E. Seidel (Mechithar, p. 121). The latter refers to the man- 

6 Thus also Spiegel, Eranische Altertumskunde, Vol. I, p. 259. 

444 Sino-Iranica 

call it Batiec Indi, which means 'melon of India/ and Avicenna so calls 
it in many places." 1 Nor does Persian herbuz, 2 Middle Persian harbojlna 
or xarbuzak (literally, " donkey-cucumber ") favor the assumption of 
an indigenous origin. VAmbery 3 argues that Turkish karpuz or harbuz 
is derived from the Persian, and that accordingly the fruit hails from 
Persia, though the opposite standpoint would seem to be equally 
justifiable, and the above interpretation may be no more than the 
outcome of a popular etymology. But Vamb6ry, after all, may be right; 
at least, by accepting his theory it would be comparatively easy to 
account for the migration of the water-melon. In this case, Persia 
would be the starting-point from which it spread to the Turks of Central 
Asia and finally to China. 4 A philological argument may support the 
opinion that the Turkish word was derived from Persia: besides the 
forms with initial guttural, we meet an alternation with initial dental, 
due to phonetic dissimilation. The Uigur, as we know from the Uigur- 
Chinese vocabulary, had the word as karpuz; but the Mongols term the 
water-melon tarbus. Likewise in TurkI we have tarbuz, but also qarpuz. 
This alternation is not Mongol-Turkish, but must have pre-existed in 
Persian, as we have tarambuja in Neo-Sanskrit, and in Hindustani 
there is xarbuza and tarbuza (also tarbuz and tarmus), and correspondingly 
tarbuz in West-Tibetan. In Pu§tu, the language of the Afghans, we 
have tarbuja in the sense of "water-melon," and xarbuja designating 
various kinds of musk-melon. 6 Through Turkish mediation the same 
word reached the Slavs (Russian arbUz, 6 Bulgarian karpHz, Polish 
arbuz, garbuz, harbuz) and Byzantines (Greek mpTrovoia) , and Turkish 
tribes appear to have been active in disseminating the fruit east and 

It would therefore be plausible also that, as stated by Joret, 7 the 
fruit may have been propagated from Iran to India, although the 
date of this importation is unknown. From Indian sources, on the other 
hand, nothing is to be found that would indicate any great antiquity of 
the cultivation of this species. Of the alleged Sanskrit word chayapula, 

1 C. Markham, Colloquies by Garcia da Orta, p. 304. 

2 From which Armenian xarpzag is derived. 

8 Primitive Cultur des turko-tatarischen Volkes, pp. 217-218. 

4 Vamb^ry, of course, is wrong in designating Persia and India as the mother- 
country of this cultivation. The mother-country was ancient Egypt or Africa in 
a wider sense. 

8 H. W. Bellew, Report on the Yusufzais, p. 255 (Lahore, 1864). 

6 In the dialects of northern Persia we also find such forms as arhuz and arhoz 
(J. de Morgan, Mission en Perse, Vol. V, p. 212). 

7 Plantes dans l'antiquite\ Vol. II, p. 252. 

The Water-Melon 445 

which A. de Candolle introduces as evidence for the early diffusion 
of the cultivation into Asia, I cannot find any trace. The Sanskrit 
designations of the water-melon, ndfdmra ("mango of the Nata"?), 
godumba, tarambuja, sedu, are of recent origin and solely to be found in 
the lexicographers; while others, like kalinga (Benincasa cerifera), orig- 
inally refer to other cucurbitaceous plants. Watt gives only modern 
vernacular names. 

Chinese si kwa has been equated with Greek cinba by Hirth, 1 who 
arbitrarily assigns to the latter the meaning "water-melon." This 
philological achievement has been adopted by Giles in his Chinese 
Dictionary (No. 6281). The Greek word, however, refers only to the 
cucumber, and the water-melon remained unknown to the Greeks of 
ancient times. 2 A late Greek designation for the fruit possibly is ire-jruv, 
which appears only in Hippocrates. 3 A. de Candolle 4 justly remarked 
that the absence of an ancient Greek name which may with certainty 
be attributed to this species seems to show that it was introduced into 
the Graeco-Roman world about the beginning of the Christian era. 
The Middle and Modern Greek word xwrov£a or Kapirovaia, derived 
from Persian or Turkish, plainly indicates the way in which the By- 
zantine world became acquainted with the water-melon. There is, 
further, no evidence that the Greek word (tikvcl ever penetrated into 
Asia and reached those peoples (Uigur, Kitan, Jur&) whom the Chinese 
make responsible for the transmission of the water-melon. The Chinese 
term is not a transcription, but has the literal meaning "western melon"; 
and the "west" implied by this term does not stretch as far as Greece, but, 
as is plainly stated in the Wu tai H, merely alludes to the fact that the 
fruit was produced in Turkistan. Si kwa is simply an abbreviation 
for Si yii kwa W fi£ ift; that is, "melon of Turkistan." 5 

According to the Yamato-honzo 6 of 1709, water-melons were first 
introduced into Japan in the period Kwan-ei (1624-44). 

1 Fremde Einnusse in der chinesischen Kunst, p. 17. 

2 A. de Candolle, G6ographie botanique, p. 909. 

8 Even this problematic interpretation is rejected by L. Leclerc (Traite - des 
simples, Vol. I, p. 239), who identifies the Greek word with the common gourd. 
Leclerc's controversy with A. de Candolle should be carefully perused by those 
who are interested in the history of the melon family. 

* Origin of Cultivated Plants, p. 264. 

5 Illustrations of Chinese water-melon fields may be seen in F. H. King, Farm- 
ers of Forty Centuries, pp. 282, 283. 

• Ch. 8, p. 3. 


50. In regard to the fenugreek (Trigonella foenum-graecuni, French 
fenugrec), Chinese hu-lu-pa (Japanese koroha) t$ M. Br, Stuart 1 states 
without further comment that the seeds of this leguminous plant were 
introduced into the southern provinces of China from some foreign 
country. But Bretschneider 2 had correctly identified the Chinese 
name with Arabic hulba (xulba). The plant is first mentioned in the 
Pen ts'ao of the Kia-yu period (a.d. 1056-64) of the Sung dynasty, 
where the author, Can Yii-si # M £§, says that it grows in the prov- 
inces of Kwan-tun and Kwei-cou, and that, according to some, the 
species of Lin-nan represents the seeds of the foreign lo-po (Raphanus 
sativus), but that this point has not yet been investigated. Su Sun, 
in his T x u kin pen ts'ao, states that "the habitat of the plant is at present 
in Kwah-tun, and that in the opinion of some the seeds came from 
Hai-nan and other barbarians; passengers arriving on ships planted 
the seeds in Kwan-tuii (Lin-wai), where the plant actually grows, but 
its seeds do not equal the foreign article; the seeds imported into China 
are really good." Then their employment in the pharmacopoeia is 
discussed. 3 The drug is also mentioned in the Pen ts'ao yen i. 4 

The transcription hu-lu-pa is of especial interest, because the 
element hu forms part of the transcription, but may simultaneously 
imply an allusion to the ethnic name Hu. The form of the transcription 
shows that it is post-T'ang; for under the T'ang the phonetic equiva- 
lent of the character $J was still possessed of an initial guttural, and a 
foreign element xu would then have been reproduced by a quite different 

The medical properties of the plant are set forth by Abu Mansur in 
his Persian pharmacopoeia under the name hulbat. 6 The Persian name 

1 Chinese Materia Medica, p. 442. 

2 Bot. Sin., pt. 1, p. 65. 

3 Stuart (/. c.) says wrongly that the seeds have been in use as a medicine since 
the T'ang dynasty; this, however, has been the case only since the Sung. I do not 
know of any mention of the plant under the T'ang. This negative documentary 
evidence is signally confirmed by the transcription of the name, which cannot have 
been made under the T'ang. 

* Ch. 12, p. 4 b (ed. of Lu Sin-yuan). 

6 Achundow, Abu Mansur, p. 47. Another Persian form is hulya. In Arme- 
nian it is hulba or hulbe (E. Seidel, Mechithar, p. 183). See also Leclerc, Traite" 


Fenugreek 447 

is ianbaltd, Sanbaltle in Ispahan, and lamllz in Shiraz, which appears 
in India as lamli. As is well known, the plant occurs wild in Kashmir, 
the Panjab, and in the upper Gangetic plain, and is cultivated in many 
parts of India, particularly in the higher inland provinces. The Sanskrit 
term is meihl, meihika, or methini. 1 In Greek it is /Sowcepas ("ox-horn"), 2 
Middle Greek xovXirev (from the Arabic), Neo-Greek ttjXu; Latin 
joenum graecum. 3 According to A. de Candolle, 4 the species is wild 
(besides the Panjab and Kashmir) in the deserts of Mesopotamia and 
of Persia, and in Asia Minor. John Fryer 5 enumerates it among the 
products of Persia. 6 

Another West-Asiatic plant introduced by the Arabs into China under the 
Sung is ffl ^ jl[ ya-pu-lu, first mentioned by Cou Mi ffi $? (1230-1320) as a 
poisonous plant growing several thousand li west from the countries of the Moham- 
medans (Kwei sin tsa H, sii tsi A, p. 38, ed. of Pai hai; and Ci ya fan tsa I'ao, Ch. A, 
p. 40 b, ed. of Yue ya fan ts'un Su). This name is based on Arabic yabruh or abruh 
(Persian jabruh), the mandragora or mandrake. This subject has been discussed by 
me in detail in a monograph "La Mandragore" (in French), T'oung Pao, 1917, 
pp. 1-30. 

des simples, Vol. I, p. 443. Schlimmer (Terminologie, p. 547) remarks, "L'infusion 
de la semence est un remede favori des m^decins indigenes dans les blennorhagies 
urethriques chroniques." 

1 It occurs, for instance, as a condiment in an Indian tale of King Vikramaditya 
(A. Weber, Abh. Berl. Akad., 1877, p. 67). 

2 Hippocrates; Theophrastus, Hist, plant., IV. iv, 10; or t<JXw: ibid., III. xvi, 
2; Dioscorides, II, 124. 

* Pliny, xxiv, 120. 

4 Origin of Cultivated Plants, p. 112. 

5 New Account of East India and Persia, Vol. II, p. 311. 

6 For further information see Fluckiger and Hanbury, Pharmacographia, 
P- 172. 


51. The nux-vomica or strychnine tree (Strychnos nux-vomica) 
is mentioned in the Pen ts'ao kan mu under the name # /ft W£ fan 
mu-pie ("foreign mu-pie," Momordica cochinchinensis, a cucurbitaceous 
plant), with the synonymes ^ 1! J" ma ts'ien-tse ("horse-coins," re- 
ferring to the coins on a horse's bridle, hence Japanese matin), i§f %. 
JC fi. k'u H pa tou ("pa-tou [Croton tiglium] with bitter fruits"), 1 and 
iK ik. % ] \ JC f$ hwo-H-k'o pa-tu. The latter term, apparently of foreign 
origin, has not yet been identified; and such an attempt would also 
have been futile, as there is an error in the transcription. The correct 
mode of writing the word which is given in the Co ken lu, 2 written in 
a.d. 1366, is *K & M hwo-H-la, and this is obviously a transcription of 
Persian kulla or kulula ("nux-vomica"), a name which is also current 
in India (thus in Hindustani; Bengali kutila). The second element 
pa-tu is neither Persian nor Arabic, and, in my opinion, must be ex- 
plained from Chinese pa-tou {Croton tiglium). 

The text of the Co ken lu is as follows: "As regards hwo-H-la pa-tu, 
it is a drug growing in the soil of Mohammedan countries. In appear- 
ance it is like mu-pie-tse (Momordica cochinchinensis), but smaller. It 
can cure a hundred and twenty cases; for each case there are special 
ingredients and guides." This is the earliest Chinese mention of this 
drug that I am able to trace; and as it is not yet listed in the Cen lei 
pen ts'ao of 1108, the standard work on materia medica of the Sung 
period, it is justifiable to conclude that it was introduced into China 
only in the age of the Mongols, during the fourteenth century. This is 
further evidenced by the very form of the transcription, which is in 
harmony with the rules then in vogue for writing foreign words. The 
Kwan k'iin fan p'u z cites no other source relative to the subject than 
the Pen ts'ao kan mu, which indeed appears to be the first and only 

1 This name does not mean, as asserted by Stuart (Chinese Materia Medica, 
p. 425), "bitter-seeded Persian bean." Stuart {ibid., p. 132) says that the Arabic 
name for Croton tiglium is "batoo, which was probably derived from the Chinese 
name pa tou EL l£." True it is that the Arabs are acquainted with this plant as an 
importation from China (L. Leclerc, Traits des simples, Vol. II, p. 95), but only 
under the name dend. I fail to trace a word batu in any Arabic dictionary or in Ibn 

2 Ch. 7, p. 5 b. See above, p. 386. 

3 Ch. 6, p. 7. 


Nux-Vomica 449 

Pen ts'ao to notice it. The point is emphasized that the drug serves 
for the poisoning of dogs. The plant now grows in Se-S'wan. 

The Sanskrit term for mix- vomica is kupllu, from which is derived 
Tibetan go-byi-la or go-bye-la. 1 The latter is pronounced go-ji-la, hence 
the Mongols adopted it as gojila. It is uncertain whether the Sanskrit 
name is related to Persian kulla or not. 

According to Fluckiger and H anbury, 2 the tree is indigenous to 
most parts of India, especially the coast districts, and is found in Burma, 
Siam, Cochin-China, and northern Australia. The use of the drug in") 
India, however, does not seem to be of ancient date, and possibly was J 
taught there by the Mohammedans. It is mentioned in the Persian 
pharmacopoeia of Abu Mansur (No. 113) under the Arabic name jauz 
ul-qei. 3 Schlimmer 4 gives also the terms azaragi and gatel el-kelbe, and 
observes, "Son emploi dans la paralysie est d'ancienne date, car l'auteur 
du Mexzen el-Edviyeh en parle d£ja, ajoutant en outre que la noix vo- 
mique est un remede qui change le temperament froid en temperament 
chaud; le mSme auteur recommande les cataplasmes avec sa poudre 
dans la coxalgie et dans les maladies articulaires." 

The Arabs, who say that the tree occurs only in the interior of 
Yemen, were well acquainted with the medicinal properties of the fruit. 5 
Nux-vomica is likewise known in Indo-China (Cam salain and phun 
akam, Khmer sleh, Annamese ku-Zi; the latter probably a transcription 
of kuiila)} 

The Kew Bulletin for 191 7 (p. 341) contains the following notice on 
Strychnos nux-vomica in Cochin-China: "In K. B. 1917 (pp. 184, 185), 
some evidence is given as to the occurrence of this species in Cochin- 
China in the wild state. Since the account was written a letter and a 
packet of undoubted nux-vomica seeds have been received from the 
Director, Agricultural and Commercial Services, Cochin-China, with 
the information that the seeds were obtained from trees growing wild 
in the country. H. B. M.'s Consul, Saigon, also sends the following 
information about S. nux-vomica in Cochin-China which he has received 
from Monsieur Morange, Director of the Agricultural and Commercial 

1 Cf. Loan-Words in Tibetan, No. 50 (T'oung Pao, 1916, p. 457). 

J Pharmacographia, p. 428. 

* Achundow, Abu Mansur, p. 43. 

4 Terminologie, p. 402. 

5 L. Leclerc, Traite" des simples, Vol. I, p. 380. 

6 Cf. E. Perrot and P. Hurrier, Matiere m£dicale et pharmacop£e sino- 
annamites, p. 171; the Chinese and Annamese certainly did not avail themselves 
of this drug "from time immemorial," as stated by these authors. See, further, 
C. Ford, China Review, Vol. XV, 1887, p. 220. 

45° Sino-Iranica 

Services of Cochin-China, and also a sample of the seeds obtained from 
a Chinese exporter. The tree exists in the Eastern provinces of Cochin- 
China, principally in the forests of Baria. The seeds are bought by 
Chinese from the savage tribes known as Mois, who collect them in the 
forest; the Chinese then export them to China or sell them again to 
firms exporting to Europe. The time of fruiting is in November and 
December. M. Morange considers that the tree is certainly indigenous 
in Cochin-China, and was not introduced by early traders." If the 
tree is indigenous there, it was certainly discovered there, as far as the 
Chinese are concerned, only after the Mongol period. H. Maitre 1 deals 
with the poisons used by the Moi for their arrows, and arrives at the 
conclusion that they are derived from the upas tree (Antiaris). He does 
not mention Strychnos. 

1 Les regions Moi du sud indo-chinois, pp. 119-121 (Paris, 1909). 


52. The carrot 1 (Daucus carota),hu lo-po (Japanese ninjin) ffli H ^S 
("Iranian turnip"), a native of northern Europe, was first introduced 
into China at the time of the Yuan dynasty (a.d. 1 260-1367). This is 
the opinion of Li Si-Sen, who states that the vegetable first appeared 
at the time of the Yuan from the land of the Hu; and it is likewise main- 
tained in the Kwan k'iin fan p*u 2 that the carrot first came from the 
countries beyond the frontier j& il. I know of no text that would give 
a more detailed account of its introduction or allude to the country of 
its origin. Nevertheless it is very likely that this was some Iranian 
region. Li Si-cen states that in his time it was abundantly culti- 
vated in the northern part of the country and in San-tun, likewise 
in middle China. 3 

The history of the carrot given by Watt 4 after G. Birdwood suffers 
from many defects. A fundamental error underlies the statement, 
"In fact, the evidence of cultivation would lead to the inference that 
the carrot spread from Central Asia to Europe, and if so it might be 
possible to trace the European names from the Indian and Persian." 
On the contrary, the carrot is a very ancient, indigenous European 
cultivation, which is by no means due to the Orient. Carrots have been 
found in the pile-dwellings of Robenhausen. 5 It is not to the point, either, 
that, as stated by Watt and Birdwood, "indeed the carrot seems to 
have been grown and eaten in India, while in Europe it was scarcely 
known as more than a wild plant." The Anglo-Saxons cultivated the 
carrot in their original habitat of Schleswig-Holstein at a time when, 
in my opinion, the carrot was not yet cultivated in India; and they con- 

1 From French carole, now carotte, Italian carota, Latin carota; Greek kclpo>t6v 
(in Diphilus). This word has supplanted Anglo-Saxon moru, from *morhu (Old 
High German moraha, morha; Russian morkov', Slovenian tnrkva). Regarding the 
origin of the word lo-po, cf. T'oung Pao, 1916, pp. 83-86. 

1 Ch. 4, p. 24. I 

' A designation for the carrot not yet indicated is fu {£ lo-po, derived from the 
three fu 2 {£, the three decades of the summer, extending from about the middle 
of July to the middle of August: during the first fu the seeds of the carrot are planted, 
in the second fu the carrots are pale red, in the third they are yellow {San hwa hien 
ci ^ VC M JS, Ch. 16, p. 14 b, ed. 1877). 

4 Commercial Products of India, p. 489, or Dictionary, Vol. Ill, p. 45. 

5 J. Hoops, Waldbaume und Kulturpflanzen, p. 297; G. Buschan, Vorge- 
schichtliche Botanik, p. 148. 


452 Sino-Iranica 

tinued to cultivate it in England. 1 Moreover, the carrot grows wild in 
Britain and generally in the north temperate zone of Europe and Asia, 
and no doubt represents the stock of the cultivated carrot, which can 
be developed from it in a few generations. 2 It is impossible to connect 
Anglo-Saxon moru (not mora, as in Watt) with Sanskrit mula or mulaka. 
No evidence is given for the bold assertion that "the carrot appears to 
have been regularly used in India from fairly ancient times." The only 
sources quoted are Baber's Memoirs 3 and the Ain-i Akbari, both works 
of the sixteenth century. I fail to see any proof for the alleged antiquity 
of carrot cultivation in India. There is no genuine Sanskrit word for 
this vegetable. It is incorrect that "the Sanskrit gar jam originated 
the Persian zardak and the Arabic jegar" (sic, for jezer). Boehtlingk 
gives for garjara only the meaning "kind of grass." As indicated below, 
it was the Arabs who carried the carrot to Persia in the tenth century, 
and I do not believe that it was known in India prior to that time. 
According to Watt, Daucus carota is a native of Kashmir and the western 
Himalaya at altitudes of from 5000 to 9000 feet; and throughout 
India it is cultivated by Europeans, mostly from annually imported 
seed, and by the natives from an acclimatised if not indigenous stock. 
Also N. G. Mukerji 4 observes, "The English root-crop which has a 
special value as a nourishing famine-food and fodder is the carrot. Up- 
country carrot or gajra is not such a nourishing and palatable food as 
European carrot, and of all the carrots experimented with in this 
country, the red Mediterranean variety grown at the Cawnpore Experi- 
mental Farm seems to be the best." 

W. Roxburgh 5 states that Daucus carota "is said to be a native 
of Persia; in India it is only found in a cultivated state." He gives 
two Sanskrit names, — grinjana and gargara, but his editor remarks 
that he finds no authority for these. In fact, these and Watt's alleged 
Sanskrit names are not at all Sanskrit, but merely Hindi (Hindi 
gajara) ; and this word is derived from Persian (not the Persian derived 
from Sanskrit, as alleged by Watt). The only Sanskrit terms for 
the carrot known to me are yavana ("Greek or foreign vegetable") 
and pitakanda (literally, "yellow root"), which appears only in the 
Rajanighantu, a work from the beginning of the fifteenth century. This 

1 Hoops, op. cit., p. 600. 

2 A. de Candolle, Gebgraphie botanique, p. 827. 

3 Baber ate plenty of carrots on the night (December 21, 1526) when an attempt 
was made to poison him. Cf . H. Beveridge, The Attempt to Poison Babur Padshah 
{Asiatic Review, Vol. XII, 1917, pp. 301-304). 

4 Handbook of Indian Agriculture, 2d ed., p. 304. 
6 Flora Indica, p. 270. 

The Carrot 453 

descriptive formation is sufficient to show that the cultivated carrot 
was foreign to the Hindu. Also W. Ainslie 1 justly concludes, "Carrots 
appear to have been first introduced into India from Persia." 

According to Schweinfurth, 2 Daucus carota should display a very 
peculiar form in Egypt, — a sign of ancient cultivation. This requires 
confirmation. At all events, it does not prove that the carrot was 
cultivated by the ancient Egyptians. Neither Loret nor Woenig men- 
tions it for ancient Egypt. 

In Greek the carrot is <jTa<pv\Zvos (hence Syriac istaflin). It is men- 
tioned by Theophrastus 3 and Pliny; 4 SavKos or bamov was a kind of 
carrot or parsnip growing in Crete and used in medicine; hence Neo- 
Greek t6 ba<pd ("carrot"), Spanish dauco. A. de Candolle 5 is right 
in saying that the vegetable was little cultivated by the Greeks and 
Romans, but, as agriculture was perfected, took a more important place. 

The Arabs knew a wild and a cultivated carrot, the former under 
the name nehhl or nehsel, 6 the knowledge of which was transmitted to 
them by Dioscorides, 7 the latter under the names jezer, sejanariya (in 
the dialect of Magreb zorudiya), and sabahla. 8 The Arabic word dauku 
or duqil, derived from Greek Bamos, denotes particularly the seed of the 
wild carrot. 9 

Joret 10 presumes that the carrot was known to the ancient Iranians. 
The evidence presented, however, is hardly admissible: Daucus maximus 
which grows in Western Persia is only a wild species. This botanical 
fact does not prove that the Iranians were acquainted with the culti- 
vated Daucus carota. An Iranian name for this species is not known. 
Only in the Mohammedan period does knowledge of it spring up in 
Persia; and the Persians then became acquainted with the carrot under 
the Arabic name jazar or jezer, which, however, may have been derived 
from Persian gazar (gezer). It is mentioned under the Arabic name in 
the Persian pharmacopoeia of Abu Mansur, 11 who apparently copied 
from Arabic sources. He further points out a wild species under the 

1 Materia Indica, Vol. I, p. 57. 

2 Z. /. Ethnologic, Vol. XXIII, 1891, p. 662. 

3 Hist, plant., IX. XV, 5. 

4 xx, 15. 

6 G^ographie botanique, p. 827. 

6 L. Leclerc, Traitd des simples, Vol. Ill, p. 380. 

7 Leclerc, op. tit., Vol. I, p. 353. 

8 Leclerc, ibid., and p. 367. 

9 Leclerc, ibid., p. 138. 

10 Plantes dans l'antiquite\ Vol. II, p. 66. 

11 Achundow, Abu Mansur, p. 42. 

454 Sino-Iranica 

name HaSqdqul, which, according to Achundow, is Eryngium campestre. 
It is therefore very probable that it was the Arabs who introduced the 
carrot into Persia during the tenth century. Besides gazar (gezer), 
Persian names are zardak 1 and lawandar; the latter means "beet-root" 
and "carrot." 

John Fryer, who travelled in India and Persia from 1672 to 1681, 
enumerates carrots among the roots of Persia. 2 The late arrival of the 
vegetable in Persia is signally confirmed by the Chinese tradition 
regarding its introduction under the Mongols. This is the logical 
sequence of events. 3 

Schlimmer 4 has the following note on the subject: "Ce le'gume, 
forme* en comp6te, est consider par les Persans comme un excellent 
aphrodisiaque, augmentant la quantity et ameliorant la quality du 
sperme. L'alimentation journali&re avec des carottes est fortement 
prdnee dans les hydropisies; les carottes cuites, conservees au vin aigre, 
dissiperaient Pengorgement de la rate." Only the yellow variety of 
carrot, with short, spindle-shaped roots, occurs in Fergana. 5 

1 Possibly derived from zard ("yellow"). Persian murdmun is said to denote 
a kind of wild carrot. In Osmanli the carrot is called hawuj. 

1 New Account of East India and Persia, Vol. II, p. 310 (Hakluyt Soc, 1912). 

s Regarding the Tibetan names of the carrot, see my notes in T'oung Pao, 1916, 
PP- 503-505. 

4 Terminologie, p. 176. 

6 S. Korzinski, Vegetatiqn of Turkistan (in Russian), p. 51. 


53. The Sui iw 1 mentions two aromatics or perfumes peculiar to 
K'an (Sogdiana), — kan hiah SB" 2 ^ and a-sa-na Man M II tf ^r. 
Fortunately we have a parallel text in the T'ai pHh hwan yii ki, 3 where 
the two aromatics of K'ah are given as ~H* $• ^ M §§ M ^. Hence 
it follows that the kan of the Sui Annals is no more than an abbreviation 
of kan sun, which is well known as an aromatic, and identical with the 
true spikenard furnished by Nardostachys jatamansi. It is Sanskrit 
nalada, Tibetan span spos, Persian nard or sunbul, Armenian sumbul, 
smbul, snbul, etc. 4 It is believed that the nard found by Alexander's 
soldiers in Gedrosia 5 represents the same species, while others hold 
that it was an Andropogon. 6 

The Sanskrit term nalada is found in the Fan yi min yi tsi 1 in the 
form M H $£ na-lo-t'o, *na-la-da. It is accompanied by the fanciful 
analysis nara-dhara ("held or carried by man"), because, it is said, 
people carry the fragrant flower with them in their girdles. The word 
nalada is of ancient date, for it appears in the Atharvaveda. 8 Hebrew 
nerd, Greek nardos, 9 Persian nard and nard, are derived] therefrom. 10 
Being used in the Bible, the word was carried to all European languages. 

1 Ch. 83, p. 4 b. 

2 This character is not listed in K'an-hi, but the phonetic element "jj" leaves no 
doubt that its phonetic value is kan, *kam. 

1 Ch. 183, p. 4. 

4 Abu Mansur (Achundow's translation, pp. 82, 241) mentions sunbul-i-hindi, 
the nard of India. Schlimmer (Terminologie, p. 36) identifies this name as Andro- 
pogon nardoides or Nardus indica. On the other hand, he says (p. 555) that Nar- 
dostachys or Valeriana jatamansi has not yet been found in Persia, but that it could 
be replaced in therapeutics by Valeriana sisymbrifolia, found abundantly in the 
mountains north of Teheran. 

5 Arrian, Anabasis, VI. xxn, 5. 

6 Joret, Plantes dans l'antiquite\ Vol. II, p. 648. See, further, Periplus, 48; 
and Pliny, xn, 28; Watt, Commercial Products of India, p. 792. Marco Polo 
(ed. of Yule, Vol. I, pp. 115, 272, 284) mentions spikenard as a product of Bengal, 
Java, and Sumatra. The Malayan word ndrawastu, mentioned by Yule {ibid., 
p. 287), must be connected with Sanskrit nalada. 

7 Ch. 8, p. 4 b. 

8 MacDonell and Keith, Vedic Index, Vol. I, p. 437; H. Zimmer, Altindisches 
Leben, p. 68. 

9 First mentioned by Theophrastus, IX. viii, 2, 3. 

10 See above, p. 428. 


456 Sino-Ira*tica 

According to Stuart, 1 this plant is found in the province of Yun- 
nan and on the western borders of Se-c'wan, but whether indigenous or 
transplanted is uncertain. If it should not occur in other parts of 
China, it is more likely that it came from India, especially as Yun-nan 
has of old been in contact with India and abounds in plants intro- 
duced from there. 

54. MH^ 2 *a-sar(sat)-na (Sui Su), MMM a-sie-na (Wei iu, 
Ch. 102, p. 9), is not explained. There is no doubt that this word 
represents the transcription of an Iranian, more specifically Sogdian, 
name; but the Sogdian terms for aromatics are still unknown to us. 
Hypothetical restorations of the name are *asarna, axsarna, asna. 

55. Storax, an aromatic substance (now obtained from Liquid- 
ambar orientalis; in ancient times, however, from Styrax officinalis), 
is first mentioned by Herodotus 3 as imported into Hellas by the Phoe- 
nicians. It is styled by the Chinese m- fe su-ho, *su-gap (giep), su-gab 
(Japanese sugo), being mentioned both in the Wei lio and in the Han 
Annals as a product of the Hellenistic Orient (Ta Ts'in). 4 It is said 
there, "They mix a number of aromatic substances and extract from 
them the sap by boiling, which is made into su-ho" (fe #^#I 
3& W H J$ j££ / n") . 5 It is notable that this clause opens and ends with 
the same word ho &; and it would thus not be impossible that the 
explanation is merely the result of punning on the term su-ho, which 
is doubtless the transcription of a foreign word. Aside from this sema- 
siological interpretation, we have a geographical theory expressed in the 
Kwah ci, written prior to a.d. 527, as follows: "Su-ho is produced in 
the country Ta Ts'in; according to others, in the country Su-ho. The 
natives of this country gather it and press the juice out of it to make 
it into an aromatic, fatty substance. What is sold are the sediments 

1 Chinese Materia Medica, p. 278. 

2 This character is not in K'ah-hi. It appears again on the same page of the 
Sui J« ( 4 b) in the name of the river *Na-mit $£ ^ (ZarafSan) in the kingdom 
Nan t£c, and on p. 4 a in $$ ft $£ (U, the country Na-se-po (*Na-sek-pwa; accord- 
ing to Chavannes, Documents sur les Tou-kiue, p. 146, Nakhsab or Nasaf). On 
pp. 6 b and 7 a the river Na-mit is written $ft. Cf. also Chavannes and Pelliot, 
Trait6 manich^en, pp. 58, 191. 

3 in, 107. 

4 Hou Han $u, Ch. 118, pp. 4 b — 5 a. E. H. Parker (China Review, Vol. XV, 
p. 372) indicates in an anecdote relative to Cwah-tse that he preferred the dung- 
beetle's dung-roll to a piece of storax, and infers that indirect intercourse with western 
Asia must have begun as early as the fourth century B.C., when Cwah-tse flourished. 
The source for this story is not stated, and it may very well be a product of later 

6 The Sii Han Su gives the same text with the variant, "call it su-ho." 

Aromatics — Storax 457 

of this product." 1 Nothing is known, however, in Chinese records about 
this alleged country Su-ho (*Su-gab); hence it is probable that this 
explanation is fictitious, and merely inspired by the desire to account in 
a seemingly plausible way for the mysterious foreign word. 

In the Annals of the Liang Dynasty, 2 storax is enumerated among 
the products of western India which are imported from Ta Ts'in and 
An-si (Parthia). It is explained as "the blending of various aromatic 
substances obtained by boiling their saps; it is not a product of nature." 3 
Then follows the same passage relating to the manufacture in Ta Ts'in 
as in the Kwan U; and the Lian $u winds up by saying that the product 
passes through the hands of many middlemen before reaching China, 
and loses much of its fragrancy during this process. 4 It is likewise on 
record in the same Annals that in a.d. 519 King Jayavarman of Fu-nan 
(Camboja) sent among other gifts storax to the Chinese Court. 5 

Finally, su-ho is enumerated among the products of Sasanian Persia. 6 
Judging from the commercial relations of Iran with the Hellenistic 
Orient and from the nature of the product involved, we shall not 
err in assuming that it was traded to Persia in the same manner 
as to India. 

The Chinese-Sanskrit dictionaries contain two identifications of 
the name su-ho. In the third chapter of the Yii k x ie H ti lun %k fllll ^ 
M Bi (Yogacaryabhumicastra), 7 translated in a.d. 646-647 by Hiian 
Tsan, we find the name of an aromatic in the form 2£ *§ 'H* ?M su-tu- 
lu-kia, *sut-tu-lu-kyie; that is, Sanskrit *sturuka = storax. 8 It is 
identified by Yuan Yin with what was formerly styled tf\H $£ §1 tou-lou- 
p % o, *du-lyu-bwa. 9 It is evident that the transcription su-tu-lu-kia is 
based on a form corresponding to Greek styrak-s, storak-s, styrdkion 
of the Papyri (Syriac stiraca, astorac). This equation presents the 

1 Fan yi min yi tsi, Ch. 8, p. 9; T'ai p'in yii Ian, Ch. 982, p. 1 b. 

2 Lian Im, Ch. 54, p. 7 b. 

3 The Fan yi min yi tsi, which reproduces this passage, has, "It is not a single 
(or homogeneous) substance." 

4 Cf. Hirth, China and the Roman Orient, p. 47. 

5 Cf . Pelliot, Bull, de I'Ecole frangaise, Vol. Ill, p. 270. 

9 Sui Su, Ch. 83, p. 7 b; or Cou Su, Ch. 50, p. 6. It does not follow from these 
texts, that, as assumed by Hirth (Chao Ju-kua, pp. 16, 262), su-ho or any other 
product of Persia was imported thence to China. The texts are merely descriptive 
in saying that these are products to be found in Persia. 

7 Bunyiu Nanjio, Catalogue of the Chinese Tripitaka, No. 11 70. 

8 Yi ts'ie kin yin i, Ch. 22, p. 3 b (cf. Pelliot, T'oung Pao, 1912, pp. 478-479). 
This text has been traced by me independently. I do not believe that this name is 
connected with turuska. 

9 Probably Sanskrit durvd (cf. Journal asiatique, 1918, II, pp. 21-22). 

458 Sino-Iranica 

strongest evidence for the fact that the su-ho of the Chinese designates 
the storax of the ancients. 1 

The Fan yi min yi tsi (I.e.) identifies Sanskrit p ift H" H> ffi tu-lu-se- 
kien, *tu-lu-s6t-kiam, answering to Sanskrit turuskam, with su-ho. 
In some works this identification is even ascribed to the Kwan ci of the 
sixth century (or probably earlier). In the Pien tse lei pien, 2 where the 
latter work is credited with this Sanskrit word, we find the character 
M kie, *g'ia5, in lieu of the second character lu. The term turuska 
refers to real incense (olibanum). 3 It is very unlikely that this aromatic 
was ever understood by the word su-ho, and it rather seems that some 
ill-advised adjustment has taken place here. 

T'ao Hun-kih (a.d. 451-536) relates a popular tradition that su-ho 
should be lion's ordure, adding that this is merely talk coming from 
abroad, and untrue. 4 C'en Ts'an-k'i of the eighth century states, 5 
"Lion-ordure is red or black in color; when burnt, it will dissipate the 
breath of devils; when administered, it will break stagnant blood 
and kill worms. The perfume su-ho, however, is yellow or white in 
color: thus, while the two substances are similar, they are not identical. 
People say that lion-ordure is the sap from the bark of a plant in the 
western countries brought over by the Hu. In order to make people 
prize this article, this name has been invented." This tradition as yet 
unexplained is capable of explanation. In Sanskrit, rasamala means 
"excrement," and this word has been adopted by the Javanese and 
Malayans for the designation of storax. 6 Thus this significance of the 
word may have given the incentive for the formation of that trade- 
trick, — examples of which are not lacking in our own times. 

Under the T'ang, su-ho was imported into China also from Malayan 
regions, especially from K'un-lun (in the Malayan area), described as 

1 The most important pharmacological and historical investigation of the sub- 
ject still remains the study of D. Hanbury (Science Papers, pp. 127-150), which 
no one interested in this matter should fail to read. 

2 Ch. 195, p. 8 b. 

3 Cf. Language of the Yue-chi, p. 7. 

4 He certainly does not say, as Bretschneider (Bot. Sin., pt. Ill, p. 463) wrongly 
translates, "but the foreigners assert that this is not true." Only the foreigners 
could have brought this fiction to China, as is amply confirmed by C'en Ts'ah-k'i. 
Moreover, the T'an pen lu |tf ^ ££ says straight, "This is a falsehood of the Hu." 

6 Cen lei pen ts'ao, Ch. 12, p. 52 (ed. of 1587). 

6 Bretschneider (/. c.) erroneously attributes to Garcia da Orta the statement 
that Rocamalha should be the Chinese name for the storax, and Stuart (Chinese 
Materia Medica, p. 243) naturally searched in vain for a confirmation of this name 
in Chinese books. Garcia says in fact that liquid storax is here (that is, in India) 
called Rocamalha (Markham, Colloquies, p. 63), and does not even mention China 
in this connection. 

Aromatics — Storax 459 

purple-red of color, resembling the tse Van $f W. (Pterocarpus santalinus, 
likewise ascribed to K'un-lun), strong, solid, and very fragrant. 1 This 
is Liquidambar altingiana or Altingia excelsa, a lofty deciduous tree 
growing in Java, Burma, and Assam, with a fragrant wood yielding a 
scented resin which hardens upon exposure to the air. The Arabs 
imported liquid storax during the thirteenth century to Palembang on 
Sumatra; 2 and the T'ai p'in hwan yu ki states that su-ho oil is produced 
in Annam, Palembang (San-fu-ts'i) , and in all barbarous countries, from 
a tree-resin that is employed in medicine. The Mon ki pi Van discrimi- 
nates between the solid storax of red color like a hard wood, and the 
liquid storax of glue-like consistency which is in general use. 3 

The Chinese transcription su-ho, *su-gap, has not yet been explained. 
Hirth's 4 suggestion that the Greek ori>pa£ should have been "muti- 
lated" into su-ho is hardly satisfactory, for we have to start from the 
ancient form *su-gab, which bears no resemblance to the Greek word 
save the first element. In the Papyri no name of a resin has as yet been 
discovered that could be compared to *su-gab. 6 Nor is there any such 
Semitic name (cf. Arabic lubna). In view of this situation, the question 
may be raised whether *su-gab would not rather represent an ancient 
Iranian word. This supposition, however, cannot be proved, either, in 
the present state of science. Storax appears in the Persian materia 
medica of Abu Mansur under the Arabic name mi'a* The storax called 
rose-maloes is likewise known to the Persians, and is said to be derived 

1 Cen lei pen ts'ao, I. c. This tree is mentioned in the Ku kin lu (Ch. c, p. I b, 
as a product of Fu-nan, and by Cao Zu-kwa as a variety of sandal-wood (Hirth) 
Chao Ju-kua, p. 208). Li Si-6en (Pen ts'ao kan mu, Ch. 34, p. 12) says that the 
people of Yun-nan call tse Van by a peculiar word, $f£ Sen; this is pronounced sen 
in Yun-nan, and accordingly traceable to a dialectic variation of landan, sandan, 
sandal. The Japanese term is Ulan (Matsumura, No. 2605). 

2 Hirth, Chao Ju-kua, p. 61. 

* Cf. Tien tse lei pien, Ch. 195, p. 8 b; Bretschneider, Bot. Sin., pt. Ill, 
p. 464. The Hian p'u quoted in the Pen ts'ao is the work of Ye T'ih-kwei i§| ££ SL 
not the well-known work by Huh C'u, in which the passage in question does not 
occur (see p. 2, ed. of T'an Sun is'un $u, where it is said that it is difficult to recognize 
the genuine article). For further information on liquid storax, see Hirth, Chao 
Ju-kua, p. 200. 

* Chao Ju-kua, p. 200. 

5 Muss-Arnolt (Transactions Am. Phil. Assoc, Vol. XXIII, p. 117) derives 
the Greek word from Hebrew z'ri; the Greek should have assimilated the Semitic 
loan-word to <rr6pa£ ("spike"). This is pure fantasy. The Hebrew word, moreover, 
does not relate to storax, but, according to Gesenius, denotes a balsam or resin like 
mastic (above, p. 252). The Hebrew word for Styr ax officinalis is said to be nataf 
(Exodus, xxx, 34), Septuaginta o-tok^, Vulgata stacte (E. Levesque in Diction- 
naire de la Bible, Vol. V, col. 1869-70). 

8 Achundow, Abu Mansur, p. 138. 

460 Sino-Iranica 

from a tree growing on the Island of Cabros in the Red Sea (near Kadez, 
three days' journey from Suez), the product being obtained by boiling 
the bark in salt water until it obtains the consistency of glue. 1 

56—57. The earliest notice of myrrh is contained in the Nan lou ki 
~W] iHI IS of Su Piao %■ i% (written before the fifth century a.d., but 
only preserved in extracts of later works), if we may depend on the 
Hai yao pen ts'ao, in which this extract is contained. 2 Sii Piao is made 
to say there that "the myrrh grows in the country Po-se, and is the 
pine-tree resin of that locality. In appearance it is like If ^? Sen hian 
('divine incense') and red-black in color. As to its taste, it is bitter and 
warm." Li Si-cen annotates that he is ignorant of what the product 
len hian is. In the Pei §i, myrrh is ascribed to the country Ts'ao 
(Jaguda) north of the Ts'ufi-lih (identical with the Ki-pin of the Han), 
while this product is omitted in the corresponding text of the Sui f*. 
Myrrh, further, is ascribed to Ki-pin. 3 The Cen lei pen ts'ao gives a 
crude illustration of the tree under the title mu yao of Kwan-cou (Kwah- 
tuh) , saying that the plant grows in Po-se and resembles benjoin {nan- 
si hian, p. 464), being traded in pieces of indefinite size and of black 

In regard to the subject, Li Si-cen 4 cites solely sources of the Sung 
period. He quotes K'ou Tsuh-si, author of the Pen ts'ao yen i (a.d. i i 16), 
to the effect that myrrh grows in Po-se, and comes in pieces of in- 
definite size, black in color, resembling benjoin. In the text of this work, 
as edited by Lu Sin-yuan, 5 this passage is not contained, but merely 
the medicinal properties of the drug are set forth. 6 Su Sufi observes 
that "myrrh now occurs in the countries of the Southern Sea (Nan-hai) 
and in Kwan-Cou. Root and trunk of the tree are like those of Canarium 
(kan-lan) . The leaves are green and dense. Only in the course of years 
does the tree yield a resin, which flows d6wn into the soil, and hardens into 
larger or smaller pieces resembling benjoin. They may be gathered at 
any time." 

A strange confusion occurs in the Yu yan tsa tsu, 7 where the myrtle 
(Myrtus communis) is described under its Aramaic name asa (Arabic 

1 Schlimmer, Terminologie, p. 495. 

2 Cen lei pen ts'ao, Ch. 13, p. 39; Pen ts'ao kan mu, Ch. 34, p. 17. 
8 T'ai p'in hwan yu ki, Ch. 182, p. 12 b. 

4 Pen ts'ao kan mu, I. c. 
6 Ch. 14, p. 4 b. 

6 In all probability, there is an editorial error in the edition of the Pen ts'ao 
quoted; in other editions the same text is ascribed to Ma Ci, one of the collaborators 
in the K'ai pao pen ts'ao. 

7 Ch. 18, p. 12. 

Aromatics — Myrrh 461 

as), while this section opens with the remark, "The habitat of the 
myrrh tree $1 is in Po-se." 1 It may be, however, that, as argued by 
Hirth, mu may be intended in this case to transcribe Middle and 
New Persian milrd, which means "myrtle" (not only in the Bundahisn, 
but generally). 2 Myrrh and myrtle have nothing to do with each 
other, belonging not only to different families, but even to different 
orders; nor does the myrtle yield a resin like myrrh. It therefore re- 
mains doubtful whether myrrh was known to the Chinese during the 
T'ang period; in this case, the passage cited above from the Nan lou 
ki (like many another text from this work) must be regarded as an 
anachronism. Cao Zu-kwa gives the correct information that myrrh 
is produced on the Berbera coast of East Africa and on the Hadramaut 
littoral of Arabia; he has also left a fairly correct description of how the 
resin is obtained. 3 

Li Si-cen 4 thinks that the transcription « or ^c represents a Sanskrit 
word. This, of course, is erroneous: myrrh is not an Indian product, 
and is only imported into India from the Somali coast of Africa and from 
Arabia. The former Chinese character answers to ancient *mut or 
*mur; the latter, to *mwat, mwar, or mar. The former no doubt repre- 
sents attempts at reproducing the Semito-Persian name, — Hebrew 
mor, Aramaic mura, Arabic murr, Persian mor (Greek afxvpa, vp,vpov, 
nOpov, Latin myrrha). 5 

Whether the Chinese transcribed the Arabic or Persian form, re- 
mains uncertain: if the transcription should really appear as late as 
the age of the Sung, it is more probable that the Arabic yielded the 
prototype; but if it can be carried back to the T'ang or earlier, the 
assumption is in favor of Iranian speech. 

1 Cf. Hirth, Journal Am. Or. Soc, Vol. XXX, p. 20. Owing to a curious mis- 
conception, the article of the Yu yah tsa tsu has been placed under mi hian 3§J ^ 
("gharu-wood") in the Pen ts'ao kan mu (Ch. 34, p. 10 b), for mu $j£ hian is wrongly- 
supposed to be a synonyme of mi hian. 

8 Another New-Persian word for this plant is ariiba or anltd. In late Avestan 
it is multemela (Bartholomae, Altiran. W&rt., col. 1189). I do not believe that the 
Persian word and Armenian murt are derived from Greek nvpolvi) (Schrader in 
Hehn, Kulturpflanzen, p. 238) or from Greek ubpros (Noldeke, Persische Studien, 
II, P- 43). 

3 Hirth, Chau Ju-kua, p. 197. 

4 Pen ts'ao kah mu, Ch. 34, p. 17. 

6 Pliny, xii, 34-35; Leclerc, Traits des simples, Vol. Ill, p. 300; V. Loret, 
Flore pharaonique, p. 95. The transcription *mwat appears to transcribe Javanese 
and Bali madu ("myrrh"; Malayan manisan lebah). In an Uigur text translated 
from Sogdian or Syriac appears the word zmurna or zmuran ("myrrh"), connected 
with the Greek word (F. W. K. Miller, Uigurica, pp. 5-7). 

462 Sino-Iranica 

Theophrastus 1 mentions in the country Aria a "thorn" on which 
is found a gum resembling myrrh in appearance and odor, and this 
drops when the sun shines on it. Strabo 2 affirms that Gedrosia produced 
aromatics, particularly nard and myrrh, in such quantity that Alex- 
ander's army used them, on the march, for tent-coverings and beds, 
and thus breathed an air full of odors and more salubrious. Modern 
botanists, however, have failed to find these plants in Gedrosia or any 
other region of Iran; 3 and the Iranian myrrh of the ancients, in all 
probability, represents a different species of Balsamodendron (perhaps 
B. pubescens or B. mukul). According to W. Geiger, 4 Balsamodendron 
mukul is called in Baluci bdd, bob, or boz, a word which simply means 
"odor, aroma." It is a descendant of Avestan baoiSi, which we find in 
Pahlavi as bdd, box, Sogdian/ra/SoScm, /3o5a, New Persian bol, bo (Ossetic 
bud, "incense"). 5 

It is noteworthy also that the ancient Chinese accounts of Sasanian 
Persia do not make mention of myrrh. The botanical evidence being 
taken into due consideration, it appears more than doubtful that 
the statement of the Nan Zou ki, Yu yah tsa tsu, K K ai pao pen ts'ao, and 
Ceh lei pen ts'ao, that the myrrh-tree grows in Po-se, can be referred to 
the Iranian Po-se. True it is, the tree does not occur, either, in the 
Malayan area; but, since the product was evidently traded to China by 
way of Malaysia, the opinion might gain ground among the Chinese 
that the home of the article was the Malayan Po-se. 

The Japanese style the myrrh mirura, which is merely a modern 
transcription of "myrrha." 6 

58. TsHhmu hiah'ft'M^t ("dark-wood aromatic") is attributed 
to Sasanian Persia. 7 What this substance was, is not explained; and 
merely from the fact that the name in question, as well as mu hiah 
fcHIr ("tree aromatic") and mi hiah %£%t, usually refer to costus 
root or putchuck (also pachak), we may infer that the Persian aromatic 
was of a similar character. Thus it is assumed by Hirth; 8 but the 
matter remains somewhat hypothetical. The Chinese term, indeed, has 

1 Hist, plant., IV. iv, 13. 

2 XV. 11, 3. 

* C. Joret, Plantes dans l'antiquite\ Vol. I, p. 48. 
4 Etymologie des Balu&, p. 46. 

6 In regard to the use of incense on the part of the Manichaeans, see Chavannes 
and Pelliot, Traits manicheen, pp. 302-303, 311. 

• J. Matsumura, Shokubutsu mei-i, No. 458. 

7 Wei $u, Ch. 102, p. 5 b; Sui Su, Ch. 83, p. 7 b. 

8 Chau Ju-kua, p. 221. Putchuck is not the root of Aucklandia costus, but of 
Saussurea lappa (see Watt, Commercial Products of India, p. 980). 


no botanical value, being merely a commercial label covering different 
roots from most diverse regions. If Cao Zu-kwa compares the putchuck- 
yielding plant with Luffa cylindrica, a Cucurbitacea of southern China, 
with which he compares also the cardamom, it is perfectly clear that he 
does not visualize the genuine costus-root of Saussurea lappa, a tall, 
stout herb, indigenous to the moist, open slopes surrounding the valley 
of Kashmir, at an elevation of eight or nine thousand feet. If he further 
states that the product is found in Hadramaut and on the Somali coast, 
it is, in my opinion, not logical to reject this as "wrong," for a product 
of the name mu Man was certainly known in the China of his time 
from that region. And why not? Also Dioscorides mentions an Arabian 
costus, which is white and odoriferous and of the best quality; besides, 
he has an Indian costus, black and smooth, and a Syrian variety of wax 
color, dusky, and of strong odor. It is obvious that these three articles 
correspond to the roots of three distinct species, which have certain 
properties in common; and it has justly been doubted that the modern 
costus is the same thing as that of the ancients. The Arabs have 
adopted the nomenclature of Dioscorides. 1 The Sheikh Daud dis- 
tinguishes an Indian species, white; a black one from China; and a red, 
heavy one, adding that it is said to be a tree of the kind of Agallochum. 
Nearly everywhere in Asia have been found aromatic roots which in 
one way or another correspond to the properties of the Indian kutfha. 
Thus in Tibet and Mongolia the latter is adjusted with the genus Inula; 
and the Tibetan word ru-rta, originally referring to an Inula, was 
adopted by the Buddhist translators as a rendering of Sanskrit kutfha. 2 
In the same manner, the Chinese term mu Man formerly denoted an 
indigenous plant of Yun-nan, which, according to the ancient work 
Pie lu, grew in the mountain-valleys of Yun-S'an. 3 The correctness of 
this tradition is confirmed by the Man §u, which mentions a mountain- 
range, three days' journey south of Yuh-6'afi, by name Ts'in-mu-hian 
("Dark-Wood Aromatic"), and owing its name to the great abundance 
of this root. 4 The Man iw, further, extends its occurrence to the country 

1 Leclerc, Trait6 des simples, Vol. Ill, pp. 85-86. 

2 H. Laufer, Beitrage zur Kenntnis der tibetischen Medicin, p. 61. 

1 Also Wu K'i-tsun (Ci wu min Si t'u k'ao, Ch. 25, p. 11) observes correctly that 
this species is not the putchuck coming from the foreign barbarians. His three 
illustrations, putchuck from Hai-cou in Kiah-su, from Kwah-tuh, and from C'u-cou 
in Nan-hwi, are reproduced from the T'u Su tsi I'en (XX, Ch. 117), and represent 
three distinct plants. 

4 The Tien hai yu hen ii (Ch. 3, p. 1 ; see above, p. 228) states that mu Man is 
produced in the native district C'6-li $i£^, formerly called C'an-li |H M» 
of Yun-nan. 

464 Sino-Iranica 

K'un-lun of the Southern Sea; 1 and Su Kun of the T'ang says that, of 
the two kinds of mu-hian (known to him), that of K'un-lun is the best, 
while that from the West Lake near Hah-cou is not good. 2 In the time 
of T'ao Hun-kin (a.d. 451-536) the root was no longer brought from 
Yuh-c'an; but the bulk of it was imported on foreign ships, with the 
report that it came from Ta Ts'in (the Hellenistic Orient), 3 — hence 
presumably the same article as the Arabian or Syrian costus of Dios- 
corides. The Nan fan ts'ao mu Iwan is cited by Cen Kwan of the seventh 
century as saying that the root is produced in India, being the product 
of an herbaceous plant and of the appearance of licorice. The same 
text is ascribed to the Nan Zou i wu U of the third century in the T'ai 
pHn yii Ian* while the Kwan U attributes the product to Kiao-cou 
(Tonking) and India. A different description of the plant is again given 
by Su Sufi. Thus it is no wonder that the specimens from China 
submitted for identification have proved to be from different plants, 
as Aplotaxis auriculata, Aristolochia kaempferi, Rosa banksia, etc. 5 If, 
accordingly, costus (to use this general term) was found not only in 
India and Kashmir, but also in Arabia, Syria, Tibet, Mongolia, China, 
and Malacca, it is equally possible also that Persia had a costus of her 
own or imported it from Syria as well as from India. 6 This is a question 
which cannot be decided with certainty. The linguistic evidence is 
inconclusive, for the New-Persian kust is an Arabic loan-word, the 
latter, of course, being traceable to Sanskrit kustha, which has obtained 
a world-wide propagation. 7 Like so many other examples in the his- 
tory of commerce, this case illustrates the unwillingness of the world 
to tolerate monopolies for any length of time. The real costus was 
peculiar (and still is) to Kashmir, but everywhere attempts were con- 
stantly made to trace equivalents or substitutes. The trade-mark 
remained the same, while the article was subjected to changes. 

59. Under the term nan (or an) -si hiah $c ® ^r the Chinese have 

1 Pelliot, Bull, de VEcole frangaise, Vol. IV, p. 226. 

2 The attribution of the root to K'un-lun is not fiction, for this tradition is 
confirmed by Garcia da Orta, who localizes pucho on Malacca, whence it is exported 
to China. 

3 This text is doubtless authentic; it is already recorded in the T'ai p'in yii Ian 
(Ch. 991, p. 11). 

4 Ch. 982, p. 3. 

5 Hanbury, Science Papers, p. 257; Stuart, Chinese Materia Medica, p. 43. 

6 In the sixteenth century, as we learn from Garcia (Markham, Colloquies, 
p. 150), costus was shipped from India to Ormuz, and thence carried to Persia and 
Khorasan; it was also brought into Persia and Arabia by way of Aden. 

7 In Tokharian it is found in the form ka$$u (S. Levi, Journal asiatique, 1911, 
II, p. 138). 

Aromatics — Styrax benjoin 465 

combined two different aromatics, — an ancient product of Iranian 
regions, as yet unidentified; and the benjoin yielded by the Styrax 
benjoin, a small tree of the Malay Archipelago. 1 It is necessary to dis- 
criminate sharply between the two, and to understand that the ancient 
term originally relating to an Iranian aromatic, when the Iranian im- 
portation had ceased, was subsequently transferred to the Malayan 
article, possibly on account of some outward resemblance of the two, 
but that the two substances have no botanical and historical inter- 
relation. The attempt of Cao Zu-kwa to establish a connection between 
the two, and to conjecture that the name is derived from An-si (Parthia) , 
but that the article was imported by way of San-fo-ts'i (Palembang on 
Sumatra), 2 must be regarded as unfounded; for the question is not of 
an importation from Parthia or Persia to Sumatra, but it is the native 
product of 'a plant actually growing in Sumatra, in Borneo, and other 
Malayan islands. 3 The product is called in Malayan kaminan (Garcia: 
cominham), Javanese menan, Sunda minan. The duplicity of the article 
and the sameness of the term have naturally caused a great deal of 
confusion among Chinese authors, and perhaps no less among European 
writers. At least, the subject has not yet been presented clearly, and 
least of all by Bretschneider. 4 

According to Su Kun, nan-si Man is produced among the Western 
2un IS 2& (Si-2un), — a vague term, which may allude to Iranians 
(p. 203). Li Sim, in his Hai yao pen ts'ao, written in the second half of 
the eighth century, states that the plant grows in Nan-hai ("Southern 
Sea"; that is, the Archipelago) and in the country Po-se. The co- 
ordination with Nan-hai renders it probable that he hints at the 
Malayan Po-se rather than at Persia, the more so, as Li Si-Sen himself 
states that the plant now occurs in Annam, Sumatra, and all foreign 
countries. 5 The reason why the term nan-si was applied to the Malayan 

1 The word "benjoin" is a corruption of Arabic lubdnjdun ("incense of Java"; 
that is, Sumatra of the Arabs). The Portuguese made of this benzawi, and further 
beijoim, benjoim (in Vasco da Gama and Duarte Barbosa); Spanish benjui, menjui; 
Italian belzuino, belguino; French benjoin. Cf. R. Dozy and W. H. Engelmann, 
Glossaire des mots espagnols et portugais denves de l'arabe, p. 239; S. R. Dalgado, 
Infiuencia do vocabulario portugufis, p. 27. 

2 Hirth, Chao Ju-kua, p. 201. 

8 According to Garcia (C. Markham, Colloquies, p. 49), benjoin is only known 
in Sumatra and Siam. According to F. Pyrard (Vol. II, p. 360, ed. of Hakluyt 
Society), who travelled from 1601 to 1610, it is chiefly produced in Malacca and 

4 Bot. Sin., pt. Ill, No. 313. 

5 As the Malayan product does not fall within the scope of the present in- 
vestigation, this subject is not pursued further here (see Hirth, Chau Ju-kua, 
pp. 201-202). In Bretschneider's translation of this matter, based on the unreliable 

466 Sino-Iranica 

product may be explained from the fact that to the south-west of 
China, west of the Irawaddy, there was a city Nan-si :£: ®, mentioned 
in the Itinerary of Kia Tan and in the Man £w of the T'ang period. 1 
The exact location of this place is not ascertained. Perhaps this or 
another locality of an identical name lent its name to the product; but 
this remains for the present a mere hypothesis. The Tien hai yii hen Zi 2 
states that nan-si is produced in the native district Pa-po ta-tien 
A "S* 3*C &J ± H, formerly called A W #t *§ i&, of Yun-nan. 

The Yu yah tsa tsu 3 contains the following account: "The tree 
furnishing the nan-si aromatic is produced in the country Po-se. 4 In 
Po-se it is termed pH-sie %$ ffi tree ('tree warding off evil influences')- 5 
The tree grows to a height of thirty feet, and has a bark of a yellow-black 
color. The leaves are oblong, 6 and remain green throughout the winter. 
It flowers in the second month. The blossoms are yellow. The heart 
of the flower is somewhat greenish (or bluish). It does not form fruit. 
On scraping the tree-bark, the gum appears like syrup, which is called 
nan-si aromatic. In the sixth or seventh month, when this substance 
hardens, it is fit for use as incense, which penetrates into the abode of 
the spirits and dispels all evil." Although I am not a botanist, I hardly 
believe that this description could be referred to Styrax benjoin. This 
genus consists only of small trees, which never reach a height of thirty 
feet; and its flowers are white, not yellow. Moreover, I am not con- 
vinced that we face here any Persian plant, but I think that the Po-se 
of the Yu yah tsa tsu, as in some other cases, hints at the Malayan 
Po-se. 7 

text of the Pen ts'ao, occurs a curious misunderstanding. The sentence j^ jJl Uti 
ft M< ^ $i !ltr i s rendered by him, "By burning the true an-si hiang incense 
rats can be allured (?)." The interrogation-mark is his. In my opinion, this means, 
"In burning it, that kind which attracts rodents is genuine." 

x Cf. Pelliot, Bull, de I'Ecole francaise, Vol. IV, pp. 178, 371. 

2 Ch. 3, p. 1 (see above, p. 228). 

3 Ch. 18, p. 8 b. 

4 Both Bretschneider (Bot. Sin., pt. Ill, p. 466) and Hirth (Chao Ju-kua, 
p. 202) identify this Po-se with Persia, without endeavoring, however, to ascertain 
what tree is meant; and Styrax benzoin does not occur in Persia. Garcia already 
stated that benjuy (as he writes) is not found in Armenia, Syria, Africa, or Cyrene, 
but only in Sumatra and Siam. 

5 PH-sie is not the transcription of a foreign word; the ancient form *bik-dza 
would lead to neither a Persian nor a Malayan word. 

6 Bretschneider, who was a botanist, translates this clause (^ ^ $j), 
"The leaves spread out into four corners (!)." Literally it means "the leaves have 
four corners"; that is, they are rectangular or simply oblong. The phrase se len J5J 
>§S? with reference to leaves signifies "four-pointed," the points being understood as 

7 See the following chapter on this subject. 

Aromatics — Styrax benjoin 467 

An identification of nan-si to which Pelliot 1 first called attention 
is given in the Chinese-Sanskrit dictionary Fan yi min yi tsi, 2 where it is 
equated with Sanskrit guggula. This term refers to the gum-resin ob- 
tained from Boswellia serrata and the produce of Balsamodendron mukul, 
or Commiphora roxburghii, the bdellion of the Greeks. 3 Perhaps also 
other Balsamodendrons are involved; and it should be borne in mind 
that Balsamodendron and Boswellia are two genera belonging to the 
same family, Burseraceae or Amyrideae. Pelliot is quite right in assum- 
ing that in this manner it is easier to comprehend the name nan-si hian, 
which seems to be attached to the ancient Chinese name of the Persia 
of the Arsacides. In fact, we meet on the rocks of Baluchistan two 
incense-furnishing species, Balsamodendron pubescens and B. mukul,' 1 
observed by the army of Alexander in the deserts of Gedrosia, and col- 
lected in great quantity by the Phoenician merchants who accompanied 
him. 5 

While it is thus possible that the term nan-si hian was originally 
intended to convey the significance "Parthian aromatic," we must not 
lose sight of the fact that it is not mentioned in the ancient historical 
documents relative to Parthia (An-si) and Persia (Po-se), — a singular 
situation, which must furnish food for reflection. The article is pointed 
out only as a product of Kuca in Turkistan and the Kingdom of Ts'ao 
W (Jagu4a) north of the Ts'uh-lih. 6 

Aside from the geographical explanation, the Chinese have 
attempted also a literal etymology of the term. According to Li Si-£en, 
this aromatic "wards off evil and sets at rest 5£c M> all demoniacal 
influences t§ 3ft ; hence its name. Others, however, say that nan-si is 
the name of a country." This word-for-word interpretation is decidedly 
forced and fantastic. 

1 Toung Pao, 1 9 12, p. 480. 

2 Ch. 8, p. 10 b. 

s Cf. T'oung Pao, 1914, p. 6. 

* Joret, Plantes dans l'antiquite\ Vol. II, p. 48. The former species is called in 
Balu£i bayi or bat. 

5 Ibid., p. 649. 

6 Sui $u, Ch. 83, pp. 5 b, 7 b. 


On the preceding pages reference has repeatedly been made to the 
fact that besides the Iranian Po-se $£ M , transcribing the ancient name 
Parsa, the Chinese were also acquainted with another country and 
people of the same name, and always written in like manner, the loca- 
tion of which is referred to the Southern Ocean, and which, as will be 
seen, must have belonged to the Malayan group. We have noted several 
cases in which the two Po-se are confounded by Chinese writers; and 
so it is no wonder that the confusion has been on a still larger scale 
among European sinologues, most of whom, if the Malayan Po-se is 
involved in Chinese records, have invariably mistaken it for Persia. 
It is therefore a timely task to scrutinize more closely what is really 
known about this mysterious Po-se of the Southern Sea. Unfortunately 
the Chinese have never co-ordinated the scattered notices of the south- 
ern Po-se; and none of their cyclopaedias, as far as I know, contains 
a coherent account of the subject. Even the mere fact of the duplicity 
of the name Po-se never seems to have dawned upon the minds of 
Chinese writers; at least, I have as yet failed to trace any text insisting 
on the existence of or contrasting the two Po-se. Groping my way 
along through this matter, I can hardly hope that my study of source- 
material is complete, and I feel sure that there are many other texts 
relative to the subject which have either escaped me or are not acces- 

The Malayan Po-se is mentioned in the Man iw §£ Ht (p. 43 b), 1 
written about a.d. 860 by Fan Co $&ffi, who says, "As regards the 
country P'iao H! (Burma), it is situated seventy-five days' journey 
(or two thousand It) south of the city of Yuh-6'an. 2 ... It borders on 
Po-se 9t $r and P'o-lo-men SI *■ P^ (Brahmana) ; 3 in the west, however, 
on the city Se-li fe M." It is clearly expressed in this document that 
Po-se, as known under the T'ang, was a locality somewhere contermi- 
nous with Burma, and on the mainland of Asia. 

1 Regarding this work, see Wylie, Notes on Chinese Literature, p. 40; and 
Pelliot, Bull, de VEcole frangaise, Vol. II, p. 156; Vol. IV, p. 132. 

2 In Yun-nan. The T'ai p'in hwan yti ki gives the distance of P'iao from that 
locality as 3000 It (cf. Pelliot, Bull, de I' Ecole frangaise, Vol. IV, p. 172). The text 
of the Man Su is reproduced in the same manner in the $u kien of Kwo Yiin-t'ao 
(Ch. 10, p. 10 b), written in 1236. 

3 1 do not believe that this term relates to India in general, but take it as denot- 
ing a specific country near the boundary of Burma. 


The Malayan Po-Se — Historical Notes 469 

In another passage of the Man $u (p. 29), the question is of a place 
Ta-yin-k'un ^C ££ ?L (evidently a silver-mine), not well determined, 
probably situated on the Gulf of Siam, to the south of which the people 
of the country P'o-lo-men (Brahmana), Po-se, Se-p'o (Java), P'o-ni 
(Borneo), and K'un-lun, flock together for barter. There are many 
precious stones there, and gold and musk form their valuable goods. 1 
There is no doubt that the Malayan Po-se is understood here, and not 
Persia, as has been proposed by Pelliot. 2 A similar text is found in the 
Nan i U W M i& ("Records of Southern Barbarians"), as quoted in the 
T'ai pHh yii lan, z "In Nan-cao there are people from P'o-lo-men, Po-se, 
Se-p'o (Java), P'o-ni (Borneo), K'un-lun, and of many other heretic 
tribes, meeting at one trading-mart, where pearls and precious stones in 
great number are exchanged for gold 4 and musk." This text is identical 
with that of the Man £w, save that the trading centre of this group of 
five tribes is located in the kingdom of Nan-cao (in the present province 
of Yun-nan). E. H. Parker 5 has called attention to a mention of Po-se 
in the T'ang Annals, without expressing, however, an opinion as to 
what Po-se means in this connection. In the chapter on P'iao (Bur- 
ma) it is there stated that near the capital of that country there are 
hills of sand and a barren waste which borders on Po-se and P'o-lo-men, 
— identical with the above passage of the Man Iw. 8 

In a.d. 742, a Buddhist priest from Yan-cou on the Yangtse, Kien- 
cen $£ M by name, undertook a voyage to Japan, in the course of which 
he also touched Canton in 748. In the brief abstract of his diary given 
by the Japanese scholar J. Takakusu, 7 we read, "Dans la riviere de 
Canton, il y avait d'innombrables vaissaux appartenant aux brahmanes, 
aux Persans, aux gens de Koun-loun (tribu malaise)." The text of the 
work in question is not at my disposal, but there can be no doubt that 
it contains the triad P'o-lo-men, Po-se, K'un-lun, as mentioned in the 
Man Su, and that the question is not of Brahmans, but of the country 

1 In another passage (p. 34 b) Fan Co states that musk is obtained in all moun- 
tains of Yun-c'an and Nan-cao, and that the natives use it as a means of exchange. 

2 Bull, de I'Ecole frangaise, Vol. IV, p. 287, note 2. 
* Ch. 981, p. 5 b. 

4 The text has ^ ^. I do not know what lu ("to boil") could mean in this 
connection. It is probably a wrong reading for Jt, as we have it in the text of the 
Man iu. 

4 Burma with Special Reference to Her Relations with China, p. 14 (Rangoon, 

6 This passage is not contained in the notice of P'iao in the Kiu T'an Su 
(Ch. 197, p. 7 b). 

7 Premier Congres International des Etudes d'Extrfime-Orient, p. 58 (Hanoi, 
1903) ; cf. G. Ferrand, Textes relatifs a l'Extrfime-Orient, Vol. II, p. 638. 

470 Sino-Iranica 

and people P'o-lo-men on the border of Burma, the Po-se likewise on the 
border of Burma, and the Malayan K'un-lun. In the first half of the 
eighth century, accordingly, we find the Malayan Po-se as a seafaring 
people trading with the Chinese at Canton. Consequently also the 
alleged "Persian" settlement on the south coast of Hainan, struck by 
the traveller, was a Malayan-Po-se colony. In view of tttis situation, the 
further question may be raised whether the pilgrim Yi Tsih in a.d. 671 
sought passage at Canton on a Persian ship. 1 This vessel was bound 
for Palembang on Sumatra, and sailed the Malayan waters; again, in 
my opinion, the Malayan Po-se, not the Persians, are here in question. 

The Malayan Po-se were probably known far earlier than the T'ang 
period, for they appear to have been mentioned in the Kwan li written 
before a.d. 527. In the Hian p'u & b? of Huh C'u ^ M of the Sung, 2 
this work is quoted as saying that $u hian ?L # (a kind of incense) 3 is 
the sap of a pine-tree in the country Po-se in the Southern Sea. This 
Po-se is well enough defined to exclude the Iranian Po-se, where, more- 
over, no incense is produced. 4 

The same text is also preserved in the Hai yao pen ts*ao of Li Siin of 
the eighth century, 5 in a slightly different but substantially identical 
wording: "Zu hian grows in Nan-hai [the countries of the Southern 
Sea] : it is the sap of a pine-tree in Po-se. That kind which is red like 
cherries and transparent ranks first." K*ou Tsuh-si, who wrote the 
Pen ts'ao yen i in a.d. 1116, says that the incense of the Southern Bar- 
barians (Nan Fan) is still better than that of southern India. The 
Malayan Po-se belonged to the Southern Barbarians. The fact that 
these, and not the Persians, are to be understood in the accounts relating 
to incense, is brought out with perfect lucidity by C'en C'eh ffi. ?k, 
who wrote the Pen ts'ao pie $wo ^ ^ #'J H& in a.d. 1090, and who says, 
"As regards the west, incense is produced in India (T'ien-cu); as re- 

1 Chavannes, Religieux eminents, p. 116; J. Takakusu, I-Tsing, p. xxvm. 

2 Ed. of T'an Sun ts'un Su, p. 5. 

8 Not necessarily from Boswettia, nor identical with frankincense. The above 
text says that iu hian is a kind of hiin-lu. The latter is simply a generic term for 
incense, without referring to any particular species. I strictly concur with Pelliot 
(T'oung Pao, 1912, p. 477) in regarding hiin-lu as a Chinese word, not as the tran- 
scription of a foreign word, as has been proposed. 

4 If hiin lu is enumerated in the Sui $u among the products of Persia, this means 
that incense was used there as an import-article, but it does not follow from this 
that "it was brought to China on Persian ships" (Hirth, Chau Ju-kua, p. 196). 
The "Persian ships," it seems, must be relegated to the realm of imagination. 
Only from the Mohammedan period did really Persian ships appear in the far east. 
The best instance to this effect is contained in the notes of Hwi Cao of the eighth 
century (Hirth, Journal Am. Or. Soc, 1913, p. 205). 

5 Pen ts'ao kan mu, Ch. 34, p. 16. 

The Malayan Po-Se — Historical Notes 471 

gards the south, it is produced in Po-se and other countries. That of 
the west is yellow and white in color, that of the south is purple or 
red." It follows from this text that the southern Po-se produced a kind 
of incense of their own; and it may very well be, that, as stated in the 
Kwah H, a species of pine was the source of this product. 

The Kwah li contains another interesting reference to Po-se. It 
states that the tree W ko, *ka (Quercus cuspidata), grows in the moun- 
tains and valleys of Kwan-tuh and Kwan-si, and that Po-se people use 
its timber for building boats. 1 These again are Malayan Po-se. The 
Kwah H was possibly written under the Tsin dynasty (a.d. 265-420) , 2 
and the Iranian Po-se was then unknown to China. Its name first 
reached the Chinese in a.d. 461, when an embassy from Persia arrived 
at the Court of the Wei. 3 It should be borne in mind also that Persia's 
communications with China always took place overland by way of 
Central Asia; while the Malayan Po-se had a double route for reaching 
China, either by land to Yun-nan or by sea to Canton. It would not 
be impossible that the word *ka for this species of oak, and also its 
synonyme /fv tZ mu-nu, *muk-nu, are of Malayan-Po-se origin. 

The Kiu yiiWhi^ 1&, published by Wan Ts'un 3: # in a.d. 1080, 
mentions that the inhabitants of Po-se wear a sort of cotton kerchief, 
and make their sarong (tu-man HP wk) of yellow silk. 4 

In a.d. 1 103, three countries, Burma, Po-se, and K'un-lun, presented 
white elephants and perfumes to the King of Ta-li in Yun-nan. Again, 
this is not Persia, as translated by C. Sainson. 5 Persia never had any 
relations with Yun-nan, and how the transportation of elephants from 
Persia to Yun-nan could have been accomplished is difficult to realize. 
We note that the commercial relations of these Po-se with Yun-nan, 
firmly established toward the end of the ninth century under the T'ang, 
were continued in the twelfth century under the Sung. 

In the History of the Sung Dynasty occurs an incidental mention of 
Po-se. 6 In a.d. 992 an embassy arrived in China from Java, and it is 
said that the envoys were dressed in a way similar to those of Po-se, who 

1 This passage is transmitted by Li Sun of the eighth century in his Hai yao 
pen ts'ao (Pen ts'ao kan tnu, Ch. 35 b, p. 14), who, as will be seen, mentions several 
plants and products of the Malayan Po-se. 

2 Pelliot, Bull, de VEcole francaise, Vol. IV, p. 412. 

3 Cf. Dev£ria in Centenaire de l'Ecole des Langues Orientales, p. 306. 

4 E. H. Parker, who made this text known {China Review, Vol. XIX, 1890, 
191), remarked, "It seems probable that not Persia, but one of the Borneo or 

Malacca states, such as P'o-li or P'o-lo, is meant." 

5 Histoire du Nan-tchao, p. 101 (translation of the Nan Sao ye Si, written by 
5fan Sen in 1550). 

• Sun Si, Ch. 489. 

472 Sino-Iranica 

had brought tribute before. The Javanese could hardly be expected 
to have been dressed like Persians, as rashly assumed by Groeneveldt; 1 
but they were certainly dressed like their congeners, the Malayan Po-se. 

Cou K'u-fei, in his Lin wai tai ta, 2 written in 1 178, gives the following 
description of the country Po-se: "In the South-Western Ocean there 
is the country Po-se. The inhabitants have black skin and curly hair. 
Both their arms are adorned with metal bracelets, and they wrap 
around their bodies a piece of cotton-cloth with blue patterns. There 
are no walled towns. Early in the morning, the king holds his court, 
being seated cross-legged on a bench covered with a tiger-skin, while his 
subjects standing beneath pay him homage. In going out he is carried 
in a litter (Ift 9B Iwan tou), or is astride an elephant. His retinue con- 
sists of over a hundred men, who, carrying swords and shouting (to clear 
the way), form his body-guard. They subsist on flour products, meat, 
and rice, served in porcelain dishes, and eat with their fingers." The 
same text has been reproduced by Cao Zu-kwa with a few slight changes. 
His reading that Po-se is situated "above the countries of the south- 
west" is hardly correct. 3 At all events, the geographical definition of 
the Sung authors is too vague to allow of a safe conclusion. The expres- 
sion of the Lin wai tai ta does not necessarily mean that Po-se was lo- 
cated on an island, and Hirth infers that we might expect to find it in 
or near the Malay Peninsula. However vague the above description 
may be, it leaves no doubt of the fact that the tribe in question is one of 
Malayan or Negrito stock. 

As far as I know, no mention is made of the Malayan Po-se in the 
historical and geographical texts of the Ming, but the tradition regard- 
ing that country was kept alive. In discussing the a-lo-p'o (Cassia 
fistula) of C'en Ts'ah-k'i, as noted above (p. 420), Li Si-cen annotates 
that Po-se is the name of a country of^the barbarians of the south-west 

There is some evidence extant that the language of Po-se belongs to 
the Malayan family. Tsuboi Kumazo 4 has called attention to the 
numerals of this language, as handed down in the Kodanlo (Memoirs 
of Oye), a Japanese work from the beginning of the twelfth century. 
These are given in Japanese transcription as follows: — 

1 sasaa, sasaka 6 namu 20 toaro 

2 toa 7 tokti, tomu 30 akaro, akafuro 

3 naka, maka 8 jembira, or gemmira 40 hiha-furo 

4 namuha (nampa) 9 sa-i-bira, or sa-i-mi-ra 100 sasarato, sasaratu 

5 rima (lima) 10 sararo, or lararo 1000 sasaho, sasahu 

1 Notes on the Malay Archipelago, p. 144. 

2 Ch. 3, p. 6 b. 

3 Ch. A, p. 33 b; Hirth's translation, p. 152. 

4 Actes du Douzieme Congres des Orientalistes, Rome 1899, Vol. II, p. 121. 

The Malayan Po-Se — Language 473 

Florenz has correctly recognized in this series the numerals of a Malayan 
language, though they cannot throughout be identified (and this could 
hardly be expected) with the numerals of any known dialect. Various 
Malayan languages must be recruited for identification, and some forms 
even then remain obscure. The numeral 1 corresponds to Malayan sa, 
satu; 2 to dua; 4 to ampat; 5 to lima; 6 to namu; 7 to tujoh; 9 to sembilan; 
10 to sa-puloh. The numeral 20 is composed of toa 2 and ro 10 (Malayan 
puloh) ; 30 aka ( = naka, 3) and ro orfuro 10. The numeral 100 is formed 
of sasa 1 and rato = Malayan -ratus. 

Two Po-se words are cited in the Yu yah tsa tsu, 1 which, as formerly 
pointed out by me, cannot be Persian, but betray a Malayan origin. 2 
There it is said that the Po-se designate ivory as & m pai-han, and 
rhinoceros-horn as M m hei-han. The former corresponds to ancient 
*bak-am; the latter, to *hak-am or *het-am. The latter answers 
exactly to Jarai hotam, Bisaya itom, Tagalog Itim, Javanese item, 
Makasar etah, Cam hutam (hatam or hutum), Malayan hitam, all mean- 
ing "black." 3 The former word is not related to the series putih, puteh, 
as I was previously inclined to assume, but to the group: Cam bauh, 
boh, or bhuh; Senoi biug, other forms in the Sakei and Semang lan- 
guages of Malakka biok, biak, bieg, begitik, bekuh, bekog;* Alfur, Boloven, 
Kon tu, Kaseng, Lave, and Niah bok, Sedeng r'dboh, Stieng bok 
("white ") ; Bahnar bak (Mon bu)} It almost seems, therefore, as if the 
speech of Po-se bears some relationship to the languages of the tribes 
of Malacca. The Po-se distinguished rhinoceros-horn and ivory as 
"black" and "white." However meagre the linguistic material may be, 
it reveals, at any rate, Malayan affinities, and explodes Bretschneider's 
theory 6 that the Po-se of the Archipelago, alleged to have been on 
Sumatra, owes its origin to the fact that "the Persians carried on a 
great trade with Sumatra, and probably had colonies there." This is an 
unfounded speculation, justly rejected also by G. E. Gerini: 7 these 
Po-se were not Persians, but Malayans. 

The Po-se question has been studied to some extent by G. E. 
Gerini, 8 who suggests its probable identity with the Vasu state located 
by the Bhagavata Purana in Kugadvipa, and who thinks it may be 

1 Ch. 16, p. 14. 

2 Chinese Clay Figures, p. 145. 

5 Cf. Cabaton and Aymonier, Dictionnaire c"am-francais, p. 503. 

4 P. Schmidt, Bijdragen tot de Taat-, Land- en Volkenkunde, Vol. VIII, 1901 , 
p. 420. 

5 Ibid., p. 344. 

6 Knowledge possessed by the Chinese of the Arabs, p. 16. 

7 Researches on Ptolemy's Geography of Eastern Asia, p. 471. 

8 Ibid., p. 682. 

474 Sino-Iranica 

Lambesi; i.e., Besi or Basi (lam meaning "village"), a petty state on 
the west coast of Sumatra immediately below Acheh, upon which it 
borders. This identification is impossible, first of all, for phonetic reasons : 
Chinese po $fc was never possessed of an ancient labial sonant, but 
solely of a labial surd (*pwa). 1 

Tsuboi Kumazo 2 regards Po-se as a transcription of Pasi, Pasei, 
Pasay, Pazze, or Pacem, a port situated on northern Sumatra near the 
Diamond Cape, which subsequently vied in wealth with Majapahit 
and Malacca, and called Basma by Marco Polo. 3 

C. O. Blagden 4 remarks with reference to this Po-se, "One is very 
much tempted to suppose that this stands for Pose (or Pasai) in north- 
eastern Sumatra, but I have no evidence that the place existed as early 
as 1178." If this be the case, the proposed identification is rendered 
still more difficult; for, as we have seen, Po-se appears on the horizon 
of the Chinese as early as from the seventh to the ninth century under the 
T'ang, and probably even at an earlier date. The only text that gives 
us an approximate clew to the geographical location of Po-se is the 
Man $u; and I should think that all we can do under the circumstances, 
or until new sources come to light, is to adhere to this definition; 
that is, as far as the T'ang period is concerned. Judging from the 
movements of Malayan tribes, it would not be impossible that, in the 
age of the Sung, the Po-se had extended their seats from the mainland 
to the islands of the Archipelago, but I am not prepared for the present 
either to accept or to reject the theory of their settlement on Sumatra 
under the Sung. 

Aside from the references in historical texts, we have another class 
of documents in which the Malayan Po-se is prominent, the Pen-ts'ao 
literature and other works dealing with plants and products. I propose 
to review these notices in detail. V 

60. In regard to alum, F. P. Smith 5 stated that apart from native 
localities it is also mentioned as reaching China from Persia, K'un-lun, 

1 On p. 471 Gerini identifies Po-se with the Baslsi tribe in the more southern 
parts of the Malay Peninsula. On the other hand, it is difficult to see why Gerini 
searched for Po-se on Sumatra, as he quotes after Parker a Chinese source 
under the date a.d. 802, to the effect that near the capital of Burma there were 
hills of sand, and a barren waste which borders on Po-se and P'o-lo-men (see 
above, p. 469). 

2 Actes du Douzieme Congres des Orientalistes, Rome 1899, Vol. II, p. 92. 

8 Cf. Yule, Marco Polo, Vol. II, pp. 284-288. Regarding the kings of Pase, 
see G. Ferrand, Textes relatifs a l'Extr6me-Orient, Vol. II, pp. 666-669. 

4 Journal Royal As. Soc, 1913, p. 168. 

5 Contributions towards the Materia Medica of China, p. 10. 

The Malayan Po-Se — Alum 475 

and Ta Ts'in. J. L. Soubeiran 1 says, "L'alun, qui £tait tire" primitive- 
ment de la Perse, est aujourd'hui importe" de l'Occident." F. de Mely 2 
translates the term Pose ts'e fan by "fan violet de Perse." All this is 
wrong. Hirth 3 noted the difficulty in the case, as alum is not produced 
in Persia, but principally in Asia Minor. Pliny 4 mentions Spain, 
Egypt, Armenia, Macedonia, Pontus, and Africa as alum-producing 
countries. Hirth found in the P l ei wen yiin fu a passage from the Hai 
yao pen ts'ao, according to which Pose fan $k.M !£ ("Persian alum," 
as he translates) comes from Ta Ts'in. In his opinion, " Persian alum" 
is a misnomer, Persia denoting in this case merely the emporium from 
which the product was shipped to China. The text in question is not 
peculiar to the Hai yao pen ts'ao of the eighth century, but occurs at a 
much earlier date in the Kwan cou ki M 'J'H Ifi, an account of Kwan- 
tuh, written under the Tsih dynasty (a.d. 265-419), when the name of 
Persia was hardly known in China. This work, as quoted in the Cen 
lei pen ts'ao, h states that kin sien 4£ H fan ("alum with gold threads") 
is produced ^fe in the country Po-se, and in another paragraph that the 
white alum of Po-se {Pose pai fan) comes from Ta Ts'in. 6 The former 
statement clearly alludes to the alum discolored by impurities, as still 
found in several localities of India and Upper Burma. 7 Accordingly 
the Malayan Po-se (for this one only can come into question here) 
produced an impure kind of alum, and simultaneously was the transit 
mart for the pure white alum brought from western Asia by way of 
India to China. It is clear that, because the native alum of Po-se was 
previously known, also the West-Asiatic variety was named for Po-se. 
A parallel to the Pose fan is the K'un-lun fan, which looks like black 
mud. 8 

61. The Wu lu ^1 $ifc, written by Can Po §H $& in the beginning of 
the fourth century, contains the following text on the subject of "ant- 
lac" iyi tsi Ji &) : 9 "In the district of Ku-fuh Wi M. (in Kiu-cen, Ton- 

1 Etudes sur la matiere medicale chinoise (Min£raux), p. 2 (reprint from 
Journal de pharmacie et de chimie, 1866). 

2 Lapidaire chinois, p. 260. 

* Chinesische Studien, p. 257. 

4 xxxv, 52. 

5 Ch. 3, p. 40 b. 

6 Also in the text of the Hai yao pen ts'ao, as reproduced in the Pen ts'ao kan mu 
(Ch. 11, p. 15 b), two Po-se alums are distinguished. 

7 Watt, Commercial Products of India, p. 61. 

8 Pen ts'ao kan mu, I. c. 

9 T'ai p'in hwan yti ki, Ch. 171, p. 5. 

476 Sino-Iranica 

king) 1 there are ants living on coarse creepers. The people, on examin- 
ing the interior of the earth, can tell the presence of ants from the soil 
being freshly broken up; and they drive tree-branches into these spots, 
on which the ants will crawl up, and produce a lac that hardens into a 
solid mass." Aside from the absurd and fantastic notes of Aelian, 2 this is 
the earliest allusion to the lac-insect which is called in Annamese con 
mdi, in Khmer kandier, in Cam mil, mur, or muor? The Chinese half- 
legendary account 4 agrees strikingly with what Garcia reports as the 
Oriental lore of this wonder of nature: "I was deceived for a long 
time. For they said that in Pegu the channels of the rivers deposit mud 
into which small sticks are driven. On them are engendered very large 
ants with wings, and it is said that they deposit much lacre 5 on the 
sticks. I asked my informants whether they had seen this with their 
own eyes. As they gained money by buying rubies and selling the cloths 
of Paleam and Bengal, they replied that they had not been so idle as 
that, but that they had heard it, and it was the common fame. After- 
wards I conversed with a respectable man with an enquiring mind, who 
told me that it was a large tree with leaves like those of a plum tree, and 
that the large ants deposit the lacre on the small branches. The ants 
are engendered in mud or elsewhere. They deposit the gum on the 
tree, as a material thing, washing the branch as the bee makes honey; 
and that is the truth. The branches are pulled off the tree and put in 
the shade to dry. The gum is then taken off and put into bamboo joints, 
sometimes with the branch." 6 

In the Yu yah tsa tsu 7 we read as follows: "The tse-kuh tree ^ £# 3 
$$ has its habitat in Camboja (Cen-la), where it is called Wl H lo-k'ia, 
*lak-ka (that is, lakka, lac). 9 Further, it is produced in the country 

1 Regarding this locality, cf. H. Maspero, Etudes d'histoire d'Annam, V, p. 19 
{Bull, de I'Ecole francaise, 1918, No. 3). \ 

2 Nat. Anim., iv, 46. There is no other Greek or Latin notice of the matter. 

8 Cf. Aymonier and Cabaton (Dictionnaire cam-francais, p. 393), who trans- 
late the term "termite, pou de bois, fourmi blanche." 

4 Much more sensible, however, than that of Aelian. 

6 The Portuguese word for "lac, lacquer," the latter being traceable to lacre. 
The ending -re is unexplained. 

8 C. Markham, Colloquies, p. 241. 

7 Ch. 18, p. 9. 

8 The Pai-hai edition has erroneously the character §ffi. 

9 From Pali lakha (Sanskrit lak^a, laktaka) ; Cam lak, Khmer lak; Siamese rak 
(cf. Pallegoix, Description du royaume Thai, Vol. I, p. 144). We are thus en- 
titled to trace the presence of this Indian word in the languages of Indo-China 
to the age of the T'ang. The earliest and only classical occurrence of the word is in 
the Periplus (Ch. 6: X&wcos). Cf. also Prakrit lakka; Kawi and Javanese laka; 
Tagalog lakha. 

The Malayan Po-Se — Lac 477 

Po-sc it #f . The tree grows to a height of ten feet, with branches dense 
and luxuriant. Its leaves resemble those of the Citrus and wither 
during the winter. In the third month it flowers, the blossoms being 
white in color. It does not form fruit. When heavy fogs, dew, and 
rain moisten the branches of this tree, they produce tse-kun. The en- 
voys of the country Po-se, Wu-hai Jl M and Sa-li-sen & ^'J $c by name, 
agreed in their statement with the envoys from Camboja, who were 
a le ?uh tu wet $f ffi HP U 1 and the cramana JS ty JS jfe % Si-sa-ni- 
pa-t'o (£icanibhadra?). These said, 'Ants transport earth into the 
ends of this tree, digging nests in it; the ant-hills moistened by rain 
and dew will harden and form tse-kun. 2 That of the country K'un-lun 
is the most excellent, while that of the country Po-se ranks next.' " 3 

1 Title of a military officer. 

* "The gum-lac which comes from Pegu is the cheapest, though it is as good as 
that of other countries; what causes it to be sold cheaper is that the ants, making 
it there on the ground in heaps, which are sometimes of the size of a cask, mix with 
it a quantity of dirt" (T A vernier, Travels in India, Vol. II, p. 22). 

1 The story of lacca and the ants producing it was made known in England at 
the end of the sixteenth century. John Gerarde (The Herball or Generall Historie 
of Plantes, p. 1349, London, 1597, 1st ed; or, enlarged and amended by Thomas 
Johnson, p. 1533, London, 1633) tells it as follows: "The tree that bringeth forth 
that excrementall substance, called Lacca, both in the shops of Europe and elsewhere, 
is called of the Arabians, Persians and Turkes Loc Sutnutri, as who should say Lacca 
of Sumutra: some which have so termed it, have thought that the first plentie thereof 
came from Sumutra, but herein they have erred; for the abundant store thereof 
came from Pegu, where the inhabitants thereof do call it Lac, and others of the 
same province Tree. The history of which tree, according to that famous Herbarist 
Clusius is as followeth. There is in the countrey of Pegu and Malabar, a great tree, 
whose leaves are like them of the Plum tree, having many small twiggie branches; 
when the trunke or body of the tree waxeth olde, it rotteth in sundrie places, wherein 
do breed certaine great ants or Pismires, which continually worke and labour in the 
time of harvest and sommer, against the penurie of winter: such is the diligence 
of these Ants, or such is the nature of the tree wherein they harbour, or both, that 
they provide for their winter foode, a lumpe or masse of substance, which is of a 
crimson colour, so beautifull and so faire, as in the whole world the like cannot be 
seene, which serveth not onely to phisicall uses, but is a perfect and costly colour for 
Painters, called by us, Indian Lack. The Pismires (as I said) worke out this colour, by 
sucking the substance or matter of Lacca from the tree, as Bees do make honie and 
waxe, by sucking the matter thereof from all herbes, trees, and flowers, and the in- 
habitants of that countrie, do as diligently search for this Lacca, as we in England 
and other countries, seeke in the woods for honie; which Lacca after they have found, 
they take from the tree, and drie it into a lumpe; among which sometimes there 
come over some sticks and peeces of the tree with the wings of the Ants, which have 
fallen amongst it, as we daily see. The tree which beareth Lacca groweth in Zeilan 
and Malavar, and in other partes of the East Indies." The second edition of 1633 
has the following addition, "The Indian Lacke or Lake which is the rich colour used 
by Painters, is none of that which is used in shops, nor here figured or described by 
Clusius, wherefore our Author was much mistaken in that he here confounds together 
things so different; for this is of a resinous substance, and a faint red colour, and 
wholly unfit for Painters, but used alone and in composition to make the best hard 

478 • Sino-Iranica 

The question here is of gum-lac or stick-lac (Gummi lacca; French 
laque en bdtons), also known as kino, produced by an insect, Coccus 
or Tachardia lacca, whichlives on a large number of widely different trees, 1 
called :$$ &V or @ tse-kun or tse-keh. Under the latter name it is men- 
tioned in the "Customs of Camboja" by Cou Ta-kwan; 2 under the 
former, in the Pen ts'ao yen i? At an earlier date it occurs as ^ £$i in 
the T'an hui yao* where it is said in the notice of P'iao (Burma), that 
there the temple-halls are coated with it. In all probability, this word 
represents a transcription: Li Si-cen assigns it to the Southern Bar- 

The Po-se in the text of the Yu yan tsa tsu cannot be Persia, as is 
sufficiently evidenced by the joint arrival of the Po-se and Camboja 
envoys, and the opposition of Po-se to the Malayan K'un-lun. Without 
any doubt we have reference here to the Malayan Po-se. The product 
itself is not one of Persia, where the lac-insect is unknown. 5 It should be 
added that the Yu yan tsa tsu treats of this Po-se product along with the 
plants of the Iranian Po-se discussed on the preceding pages; and there 
is nothing to indicate that Twan C'eh-si, its author, made a distinction 
between the two homophonous names. 6 

62. The Malayan Po-se, further, produced camphor (Dryobalanops 
aromatica), as we likewise see from the Yu yan tsa tsu, 7 where the tree 

sealing wax. The other seemes to be an artificiall thing, and is of an exquisite crim- 
son colour, but of what it is, or how made, I have not as yet found any thing that 
carries any probabilitie of truth." Gerarde's information goes back to Garcia, 
whose fundamental work then was the only source for the plants and drugs 
of India. 

1 Watt, Commercial Products of India, p. 1053; not necessarily Erythrina, as 
stated by Stuart (Chinese Materia Medica, p. 489). Sir C. Markham (Colloquies, 
p. 241) says picturesquely that the resinous exudation is produced by the puncture 
of the females of the lac-insect as their common nuptial and accouchement bed, the 
seraglio of their multi-polygamous bacchabunding lord, the male Coccus lacca; 
both the males and their colonies of females live only for the time they are cease- 
lessly reproducing themselves, and as if only to dower the world with one of its 
most useful resins, and most glorious dyes, the color "lake." 

2 Pelliot, Bull, de I'Ecole francaise, Vol. II, p. 166. 

3 Ch. 14, p. 4 b (ed. of Lu Sin-yuan). 

4 Ch. 100, p. 18 b. Also Su Kuh and Li Sun of the T'ang describe the product. 

5 The word lak (Arabic) or ranglak (Persian) is derived from Indian, and 
denotes either the Indian product or the gum of Zizyphus lotus and other plants 
(Achundow, Abu Mansur, p. 265). In the seventeenth century the Dutch bought 
gum-lac in India for exportation to Persia (Tavernier, /. c). Cf. also Leclerc, 
Traite" des simples, Vol. Ill, p. 241 ; and G. Ferrand, Textes relatifs a l'ExtrSme- 
Orient, p. 340. 

6 In regard to stick-lac in Tibet, see H. Laufer, Beitrage zur Kenntnis der 
tibetischen Medicin, pp. 63-64. 

7 Ch. 18, p. 8 b. 

The Malayan Po-Sb — Lac, Camphor 479 

is ascribed to Bali §1 ^'J (P'o-li, *Bwa-li) 1 ancl to Po-se. Camphor is 
not produced in Persia; 2 and Hirth 3 is not justified in here rendering 
Po-se by Persia and commenting that camphor was brought to China 
by Persian ships. 

63 . The confusion as to the two Po-se has led Twan £ 'en- si 4 to ascribe 
the jack-fruit tree (Artocarpus integrifolia) to Persia, as would follow 
from the immediate mention of Fu-lin; but this tree grows neither in 
Persia nor in western Asia. It is a native of India, Burma, and the 
Archipelago. The mystery, however, remains as to how the author 
obtained the alleged Fu-lin name. 5 

Pepper {Piper longum), according to Su Kun of the T'ang, is a prod- 
uct of Po-se. This cannot be Persia, which does not produce pepper. 6 

In the chapter on the walnut we have noticed that the Pei hu lu, 
written about a.d. 875 by Twan Kuh-lu, mentions a wild walnut as 
growing in the country Can-pei (*Cambi, Jambi), and gathered and 
eaten by the Po-se. The Lin piao lu i, written somewhat later (between 
889 and 904), describes the same fruit as growing in Can-pi (*Cambir, 
Jambir) , and gathered by the Hu. This text is obviously based on the 
older one of the Pei hu lu; and Liu Sun, author of the Lin piao lu i, 
being under the impression that the Iranian Po-se is involved, appears 
to have substituted the term Hu for Po-se. The Iranian Po-se, however, 
is out of the question: the Persians did not consume wild walnuts; 
and, for all we know about Can-pi, it must have been some Malayan 
region. 7 I have tentatively identified the plant in question with Juglans 
cathayensis or, which is more probable, Canarium commune; possibly 
another genus is intended. As regards the situation of Can-pi (or -pei) 
and Po-se of the T'ang, much would depend on the botanical evidence. 
I doubt that any wild walnut occurs on Sumatra. 

The Hai yao pen ts'ao, written by Li Sun in the second half of the 
eighth century, and as implied by the title, describing the drugs from 

1 Its Bali name is given as [§ ^fC §| ^ ku-pu-p'o-lil, *ku-put-bwa-lwut, which 
appears to be based on a form related to the Malayan type kapor-bdrus. Cf. also 
the comments of Pelliot {T'oung Pao, 1912, pp. 474-475). 

2 Schlimmer (Terminologie, p. 98) observes, "Les auteurs indigenes persans 
recommendent le camphre de Borneo comme le meilleur. Camphre de menthe, 
provenant de la Chine, se trouve depuis peu dans le commerce en Perse." Camphor 
was imported into Slraf (W. Ouseley, Oriental Geography of Ebn Haukal, p. 133; 
G. Le Strange, Description of the Province of Fars, p. 42). 

8 Chau Ju-kua, p. 194. 

4 Yu yan tsa tsu, Ch. 18, p. 10. 

5 Cf. Hirth, Chau Ju-kua, p. 213. 

6 See above, pp. 374, 375. 

7 See the references given above on p. 268. 

480 Sino-Iranica 

the countries beyond the sea and south of China, has recorded several 
products of Po-se, which, as we have seen, must be interpreted as the 
Malayan region of this name. Such is the case with benjoin (p. 464) 
and cummin (p. 383). 

We noticed (p. 460) that the Nan lou ki and three subsequent works 
attribute myrrh to Po-se, but that this can hardly be intended for 
the Iranian Po-se, since myrrh does not occur in Persia. Here the 
Malayan Po-se is visualized, inasmuch as the trade in myrrh took its 
route from East Africa and the Hadramaut coast of Arabia by way of 
the Malay Archipelago into China, and thus led the Chinese (errone- 
ously) to the belief that the tree itself grew in Malaysia. 

64. The case of aloes {Aloe vulgaris and other species) presents a 
striking analogy to that of myrrh, inasmuch as this African plant 
is also ascribed to Po-se, and a substitute for it was subsequently found 
in the Archipelago. Again it is Li Sun of the T'ang period who for the 
first time mentions its product under the name lu-wei Htl?, stating 
that it grows in the country Po-se, has the appearance of black con- 
fectionery, and is the sap of a tree. 1 Su Sun of the Sung dynasty 
observes, "At present it is only shipped to Canton. This tree grows in 
the mountain-wilderness, its sap running down like tears and coagulat- 
ing. This substance is gathered regardless of the season or month." 
Li Si-cen feels doubtful as to whether the product is that of a tree or of 
an herb ^: he points out that, according to the Ta Min i t'un Zi y 
aloes, which belongs to the class of herbs, is a product of Java, Sumatra 
(San-fu-ts'i), and other countries, and that this is contradictory to 
the data of the T'ang and Sung Pen-ts'ao. It was unknown to him, 
however, that the first author thus describing the product is Cao 
2u-kwa, 2 who indeed classifies Aloe among herbs, and derives it from 
the country Nu-fa $2 H, a dependency of the Arabs, and in another 
passage from an island off the Somali coast, evidently hinting at Socotra. 
This island is the home of the Aloe perryi, still imported into Bombay. 3 

The name lu-wei is traced by Hirth to Persian alwa. This theory is 
difficult to accept for many reasons. Nowhere is it stated that lu-wei 
is a Persian word. Li Si-Sen, who had good sense in diagnosing foreign 
words, remarks that lu-wei remains unexplained. The Chinese his- 
torical texts relative to the Iranian Po-se do not attribute to it this 
product, which, moreover, did not reach China by land, but exclusively 

1 Pen ts'ao kan mu, Ch. 34, p. 21 b. The juice of Aloe abyssinica is sold in the 
form of flat circular cakes, almost black in color. 

2 Cufan U, Ch. B, p. 11 (cf. Hirth's translation, p. 225). 

3 Regarding the history of aloes, see especially Fluckiger and Hanbury, 
Pharmacographia, p. 680. 

The Malayan Po-Se — Aloes 481 

over the maritime route to Canton. Aloes was only imported to Persia, 1 
but it is not mentioned by Abu Mansur. The two names sebr zerd 
and sebr sugutri ( = Sokotra), given by Schlimmer, 2 are of Arabic and 
comparatively modern origin; thus is likewise the alleged Persian word 
alwd. The Persians adopted it from the Arabs; and the Arabs, on their 
part, admit that their alua is a transcription of the Greek word CC.X617. 3 
We must not imagine, of course, that the Chinese, when they first re- 
ceived this product during the T'ang period, imported it themselves 
directly from the African coast or Arabia. It was traded to India, and 
from there to the Malayan Archipelago; and, as intimated by Li Sun, 
it was shipped by the Malayan Po-se to Canton. Another point over- 
looked by Hirth is that Aloe vera has been completely naturalized in 
India for a long time, although not originally a native of the country. 4 
Garcia da Orta even mentions the preparation of aloes in Cambay 
and Bengal. 5 Thus we find in India, as colloquial names for the drug, 
such forms as alia, ilva, eilya, elio, yalva, and aliva in Malayan, which 
are all traceable to the Arabic-Greek alua, alwd. This name was picked 
up by the Malayan Po-se and transmitted by them with the product to 
the Chinese, who simply eliminated the initial a of the form aluwa 
or aluwe and retained luwe. 6 Besides lu-wei, occur also the transcriptions 
18. or tH # nu or no hwi, the former in the K'ai-pao pen ts x ao of the Sung, 
perhaps suggested by the Nu-fa country or to be explained by the 
phonetic interchange of / and n. It is not intelligible to me why 
Hirth says that in the Ming dynasty lu-wei "was, as it is now, 
catechu, a product of the Acacia catechu (Sanskrit khadira)." No 
authority for this theory is cited; but this is quite impossible, as 
catechu or cutch was well known to the Chinese under the names 
er-Va or haVr-Va? 

65. A plant, fS ffi #} so-fa-mi, *suk-sa-m'it(m'ir), Japanese 
SukuSamitsu (Amomum villosum or xanthioides) , is first mentioned by Li 
Sun as "growing in the countries of the Western Sea (Si-hai) as well as 
in Si-2uh © ~$L and Po-se, much of it coming from the Nan-tun circuit 

1 W. Ouseley, Oriental Geography of Ebn Haukal, p. 133. 

* Terminologie, p. 22. 

3 Leclerc, Traite" des simples, Vol. II, p. 367. 

* G. Watt, Commercial Products of India, p. 59. 

8 C. Markham, Colloquies, p. 6. 

8 Waiters (Essays on the Chinese Language, p. 332), erroneously transcrib- 
ing lu-hui, was inclined to trace the Chinese transcription directly to the Greek 
aloe; this of course, for historical reasons, is out of the question. 

7 See Stuart, Chinese Materia Medica, p. 2; and my Loan- Words in Tibetan, 
No. 107, where the history of these words is traced. 

482 Sino-Iranica 

:2c M 3=L . " l According to Ma Ci , it grows in southern China, and, accord- 
ing to SuSun,in the marshes of Lin-nan; thus it must have been intro- 
duced between the T'ang and Sung dynasties. In regard to the name, 
which is no doubt of foreign origin, Li Si-cen observes that its significance 
is as yet unexplained. Certainly it is not Iranian, nor is it known to me 
that Amomum occurs in Persia. On the contrary, the plant has been 
discovered in Burma, Siam, Camboja, and Laos. 2 Therefore Li Sun's 
Po-se obviously relates again to the Malayan Po-se; yet his addition of 
Si-hai and Si-2un is apt to raise a strong suspicion that he himself 
confounded the two Po-se and in this case thought of Persia. I have 
not yet succeeded in tracing the foreign word on which the Chinese 
transcription is based, but feel sure that it is not Iranian. The present 
colloquial name is ts'ao ia Zen jf£ #£ C 3 

66. There is a plant styled §1 it # p'o-lo-te, *bwa-ra-tik, or §1 ft 
W) p'o-lo-lo, *bwa-ra-lak(lok, lek), not yet identified. Again our 
earliest source of information is due to Li Sun, who states, "P'o-lo-te 
grows in the countries of the Western Sea (Si-hai) and in Po-se. The 
tree resembles the Chinese willow; and its seeds, those of the castor-oil 
plant (pei-ma tse, Ricinus communis, above, p. 403) ; they are much used 
by druggists." 4 Li Si-cen regards the word as Sanskrit, and the elements 
of the transcription hint indeed at a Sanskrit name. It is evidently 
Sanskrit bhalldtaka, from which are derived Newarl pdldla, Hindustani 
belatak or bheld, Persian balddur, and Arabic beladur (Garcia : balador) . 
Other Sanskrit synonymes of this plant are aruska,bijapadapa,viravTksa, 
visdsyd, and dahana. It is mentioned in several passages of the Bower 

This is the marking-nut tree {Semecarpus anacardium, family Ana- 
cardiaceae), a genus of Indian trees found throughout the hotter parts 
of India as far east as Assam, also distributed over the Archipelago as 
far as the Philippines 5 and North Australia. It does not occur in Burma 
or Ceylon, nor in Persia or western Asia. The fleshy receptacle bear- 
ing the fruit contains a bitter and astringent substance, which is uni- 
versally used in India as a substitute for marking-ink. The Chinese 

1 Pen ts'ao kan mu, Ch. 14, p. 13 b. 

2 Stuart, Chinese Materia Medica, p. 38. Loureiro (so-xa-mi) mentions it 
for Cochin-China (Perrot and Hurrier, Mat. med. et pharmacop£e sino-annamites, 
P- 97). 

3 Ci wu min li t'u k'ao, Ch. 25, p. 72. 

4 Pen ts'ao kan mu, Ch. 35, p. 7; Cen lei pen ts'ao, Ch. 5, p. 14 b. In the latter 
work Li Sun attributes the definition "Western Sea and Po-se" to Sii Piao, author 
of the Nan lou ki. 

5 M. Blanco, Flora de Filipinas, p. 216. 

The Malayan Po-Se — Semecarpus, Psoralea 483 

say expressly that it dyes hair and mustache black. 1 It gives to cotton 
fabrics a black color, which is said to be insoluble in water, but soluble 
in alcohol. The juice of the pericarp is mixed with lime water as a 
mordant before it is used to mark cloth. In some parts of Bengal the 
fruits are regularly used as a dye for cotton cloths. 2 The fleshy cups on 
which the fruit rests, roasted in ashes, and the kernels of the nuts, are 
eaten as food. They are supposed to stimulate the mental powers, 
especially the memory. The acrid juice of the pericarp is a powerful 
vesicant, and the fruit is employed medicinally. 

In regard to the Persian-Arabic baladur, Ibn al-Baitar states express- 
ly that this is an Indian word, 5 and there is no doubt that it is derived 
from Sanskrit bhallataka. The term is also given by Abu Mansur, who 
discusses the application of the remedy. 4 The main point in this con- 
nection is that p'o-lo-te is a typical Indian plant, and that the Po-se of 
the above Chinese text cannot refer to Persia. Since the tree occurs in 
the Malayan area, however, it is reasonable to conclude that again the 
Malayan Po-se is intended. The case is analogous to the preceding 
one, and the Malayan Po-se were the mediators. At any rate, the 
transmission to China of an Indian product with a Sanskrit name by 
way of the Malayan Po-se is far more probable than by way of Persia. 
I am also led to the general conclusion that almost all Po-se products 
mentioned in the Hat yao pen ts'ao of Li Sun have reference to the 
Malayan Po-se exclusively. 

67. A drug, by the name ■fif # flit pu-ku-U (*bu-kut-tsi), identified 
with Psoralea corylifolia, is first distinctly mentioned by Ma Ci $1 JS, 
collaborator in the K'ai pao pen ts'ao (a.d. 968-976) of the Sung period, 
as growing in all districts of Lih-nan (Kwan-tuh) and K wan-si, and 
in the country Po-se. According to Ta Min ^C 9§, author of the Zi hwa 
cu kia pen ts'ao ^£f ft ^ ^ ^, published about a.d. 970, the drug 
would have been mentioned in the work Nan lou ki by Sii Piao 
(prior to the fifth century) , 5 who determined it as 1!$ M ^F hu kiu-tse, 
the "Allium odorum of the Hu." This, however, is plainly an anachro- 
nism, as neither the plant, nor the drug yielded by it, is mentioned by 
any T'ang writers, and for the first time looms up in the pharmacopoeia 
of the Sung. Su Sun, in his T'u kin pen ts'ao, observes that the plant 
now occurs abundantly on the mountain-slopes of southern China, 

1 Gen lei pen ts'ao, Ch. 5, p. 14 b. 

2 Cf. Watt, Dictionary, Vol. VI, pt. 2, p. 498. 

3 Leclerc, Traits des simples, Vol. I, pp. 162, 265. 

4 Achundow, Abu Mansur, p. 30. 

5 See above, p. 247. 

484 Sino-Iranica 

also in Ho-cou & 'M in Se-C'wan, but that the native product does not 
come up to the article imported on foreign ships. 1 Ta Mih defines the 
difference between the two by saying that the drug of the Southern 
Barbarians is red in color, while that of Kwah-tun is green. Li Si-cen 
annotates that the Hu name for the plant is §1 M Ba p x o-ku-U (*bwa- 
ku-ci, bakuCi), popularly but erroneously written Wi~$.%$> p'o-ku-H 
(*pa-ku-ci), that it is the "Allium odorum of the Hu," because the 
seeds of the two plants are similar in appearance, but that in fact it is 
not identical with the Allium growing in the land of the Hu. These 
are all the historical documents available. Stuart 2 concludes that the 
drug comes from Persia; but there is neither a Persian word bakuii, 
nor is it known that the plant {Psoralea corylijolia) exists in Persia. 
The evidence presented by the Chinese sources is not favorable, either, 
to this conclusion, for those data point to the countries south of China, 
associated in commerce with Kwah-tun. The isolated occurrence of 
the plant in a single locality of Se-6'wan is easily explained from the 
fact that a large number of immigrants from Kwah-tun have settled 
there. In fact, the word *bakuci yielded by the Chinese transcription 
is of Indian origin: it answers to Sanskrit vakucl, which indeed designates 
the same plant, Psoralea corylijolia* In Bengali and Hindustani it is 
hakul^ and bavaci, Uriya bakucl, Panjab babel, Bombay bawaci, MarathI 
bavacya or bavaci, etc. According to Watt, it is a common herbaceous 
weed found in the plains from the Himalaya through India to Ceylon. 
According to Ainslie, this is a dark brown-colored seed, about the 
size of a large pin-head, and somewhat oval-shaped; it has an aromatic, 
yet unctuous taste, and a certain degree of bitterness. The species in 
question is an annual plant, seldom rising higher than three feet; and is 
common in southern India. It has at each joint one leaf about two inches 
long, and one and a half broad; the flowers are of a pale flesh color, 
being produced on long, slender, axillary peduncles. In Annam it is 
known as hot-bo-kot-U and p'a-ko-fi. 6 It is therefore perfectly obvious 

1 According to the Gazetteer of Sen-si Province {Sen-si t'un li, Ch. 43, p. 31), 
the plant occurs in the district Si-ts'uan /£j ^. in the prefecture Hih-nan. 

2 Chinese Materia Medica, p. 359; likewise F. P. Smith (Contributions, p. 179) 
and Perrot and Hurrier (Matiere m^dicale et pharmacopee sino-annamites, 
p. 150). 

* W. Ainslie, Materia Indica, Vol. II, p. 141. 

4 This name is also given by W. Roxburgh (Flora Indica, p. 588). See, further, 
Watt, Dictionary of the Economic Products of India, Vol. VI, p. 354. 

5 Perrot and Hurrier, Mat. m6d. et pharmacopee sino-annamites, p. 150. 
According to these authors, the plant is found in the south and west of China as 
well as in Siam. Wu K'i-tsun says that physicians now utilize it to a large extent in 
lieu of cinnamon (Ci wu mih H t'u k*ao, Ch. 25, p. 65). 

The Malayan Po-Se — Ebony 485 

that the designation "Allium of the Hu" is a misnomer, and that the 
plant in question has nothing to do with the Hu in the sense of Iranians, 
nor with Persia. The Po-se of Ma Ci, referred to above, in fact repre- 
sents the Malayan Po-se. 

68. In the Pen ts'ao kan mu, a quotation is given from the Ku kin 
Zu> which is not to be found in the accessible modern editions of this 
work. The assertion is made there with reference to that work that 
ebony Mi 3C >fc is brought over on Po-se ships. It is out of the question 
that Po-se in this case could denote Persia, as erroneously assumed by 
Stuart, 1 as Persia was hardly known under that name in the fourth 
century, when the Ku kin lu was written, or is supposed to have been 
written, by Ts'ui Pao; 2 and, further, ebony is not at all a product of 
Persia. 3 Since the same work refers ebony to Kiao-cbu (Tonking), it 
may be assumed that this Po-se is intended for the Malayan Po-se; but, 
even in this case, the passage may be regarded as one of the many 
interpolations from which the Ku kin lu has suffered. 

Chinese wu-men ^li (*u-mon), "ebony" (timber of Diospyros 
ebenum and D. melanoxylon) is not a transcription of Persian dbnus, 
as proposed by Hirth. 4 There is no phonetic coincidence whatever. 
Nowhere is it stated that the Chinese word is Persian or a foreign word 
at all. There is, further, no evidence to the effect that ebony was ever 
traded from Persia to China; on the contrary, according to Chinese 
testimony, it came from Indo-China, the Archipelago, and India; 
according to Li Si-Sen, from Hai-nan, Yun-nan, and the Southern Bar- 
barians. 6 The speculation that the word had travelled east and west 
with the article from "one of the Indo-Chinese districts," is untenable; 
for the ebony of western Asia and Greece did not come from Indo- 
China, but from Africa and India. The above Chinese term is not a 
transcription at all : the second character men is simply a late substitu- 
tion of the Sung period for the older ]$C, as used in the Ku kin £u, wu wen 
meaning "black-streaked wood." In the Pen ts'ao kan mu 6 it is said 

1 Chinese Materia Medica, p. 253. 

2 Persia under the name Po-se is first mentioned in a.d. 461, on the occasion of 
an embassy sent from there to the Court of the Wei (compare above, p. 471). 

3 It was solely imported into Persia (W. Ouseley, Oriental Geography of Ebn 
Haukal, p. 133). 

* Chau Ju-kua, p. 216. 

6 The Ko ku yao lun (Ch. 8, p. 5 b; ed. of Si yin Man ts'un Su) gives Hai-nan, 
Nan-fan ("Southern Barbarians"), and Yun-nan as places of provenience, and 
adds that there is much counterfeit material, dyed artificially. The poles of the tent 
of the king of Camboja were made of ebony (Sui Su, Ch. 82, p. 3). 

6 Ch. 35 b, p. 13. 

486 Sino-Iranica 

that the character men should be pronounced in this case M man, 
that the name of the tree is 3&C ?fC (thus written in the Nan fan ts'ao mu 
ftjuah), and that the southerners, because they articulate 3C like fil, 
have substituted the latter. This is a perfectly satisfactory explanation. 
The Ku kin Zu, 1 however, has preserved a transcription in the form 
It /fC H *i-muk-i or Hf *bu (wu), which must have belonged to the 
language of Kiao-cou ^ } M (Tonking) , as the product hailed from there. 
Compare Khmer mak pen and Cam mokia ("ebony," Diospyros eben- 
aster) . 2 

Ebony was known in ancient Babylonia, combs being wrought from 
this material. 3 It is mentioned in early Egyptian inscriptions as being 
brought from the land of the Negroes on the upper Nile. Indeed, Africa 
was the chief centre that supplied the ancients with this precious wood. 4 
From Ethiopia a hundred billets of ebony were sent every third year 
as tribute to Darius, king of Persia. Ezekiel 5 alludes to the ebony of 
Tyre. The Periplus (36) mentions the shipping of ebony from Barygaza 
in India to Ommana in the Persian Gulf. Theophrastus, 6 who is the 
first to mention the ebony -tree of India, makes a distinction between two 
kinds of Indian ebony, a rare and nobler one, and a common variety of 
inferior wood. According to Pliny, 7 it was Pompey who displayed 
ebony in Rome at his triumph over Mithridates; and Solinus, who copies 
this passage, adds that it came from India, and was then shown for the 
first time. According to the same writer, ebony was solely sent from 
India, and the images of Indian gods were sometimes carved from this 
wood entirely, likewise drinking-cups. 8 Thus the ancients were ac- 
quainted with ebony as a product of Africa and India at a time when 
Indo-China was still veiled to them, nor is any reference made to the 
far east in any ancient western account of the subject. The word itself 
is of Egyptian origin: under the name jieben, ebony formed an important 
article with the country Punt. Hebrew hobntm is related to this word or 
directly borrowed from it, and Greek e'(3tvos is derived from Semitic. 
Arabic-Persian 'abniis is taken as a loan from the Greek, and Hindi 
abanusa is the descendant of abniis. 

1 Ch. c, p. 1 b. The product is described as coming from Kiao-cou, being of 
black color and veined, and also called "wood with black veins" {wu wen mu). 

2 Aymonier and Cabaton, Dictionnaire cam-frangais, p. 366. 

3 Handcock, Mesopotamian Archaeology, p. 349. 

4 Herodotus, in, 97. 

5 xxvn, 15. 

6 Hist, plant., IV. IV, 6. 

7 xn, 4, § 20. 

8 Solinus, ed. Mommsen, pp. 193, 221. 

The Malayan Po-Se and its Products 487 

It is thus obvious that the term Po-se in Chinese records demands 
great caution, and must not be blindly translated "Persia." Whenever 
it is used with reference to the Archipelago, the chances are that Persia 
is not in question. The Malayan Po-se has become a fact of historical 
significance. He who is intent on identifying this locality and people 
must not lose sight of the plants and products attributed to it. I dis- 
agree entirely with the conclusion of Hirth and Rockhill 1 that from 
the end of the fourth to the beginning of the seventh centuries all the 
products of Indo-China, Ceylon, India, and the east coast of Africa 
were classed by the Chinese as "products of Persia (Po-se)," the coun- 
try of the majority of the traders who brought these goods to China. 
This is a rather grotesque generalization, inspired by a misconception 
of the term Po-se and the Po-se texts of the Wei $u and Sui §u. The 
latter, as already emphasized, do not speak at all of any importation of 
Persian goods to China, but merely give a descriptive list of the arti- 
cles to be found in Persia. Whenever the term Po-se is prefixed to the 
name of a plant or a product, it means only one of two things, — Persia 
or the Malayan Po-se, — but this attribute is never fictitious. Not a 
single case is known to me where a specific product of Ceylon or India 
is ever characterized by the addition Po-se. 

1 Chau Ju-kua, p. 7. 


69. Brocades, that is, textiles interwoven with gold or silver threads, 
were manufactured in Iran at an early date. Gold rugs are mentioned 
in the Avesta (zaranaene upasterene, Ya§t xv, 2). Xerxes is said tc 
have presented to citizens of Abdera a tiara interwoven with gold. 1 
The historians of Alexander give frequent examples of such cloth in 
Persia. 2 Pliny, 3 speaking of gold textiles of the Romans, traces this art 
to the Attalic textures, and stamps it as an invention of the kings of 
Asia (Attalicis vero iam pridem intexitur, invento regum Asiae). 4 
The accounts of the ancients are signally confirmed by the Chinese. 

Persian brocades $t $f i& are mentioned in the Annals of the Liang as 
having been sent as tribute in a.d. 520 to the Emperor Wu from the 
country Hwa Wt* The king of Persia wore a cloak of brocade, and bro- 
^cades were manufactured in the country. 6 Textiles woven with gold 
threads ^ H£ ^ $S are expressly mentioned; 7 this term almost reads 
like a translation of Persian zar-baf (literally, "gold weaving"). 8 Per- 
sian brocades, together with cotton stuffs from An-si (Parthia) $c Hi 
6 f&t, are further mentioned at the time of the Emperor Si Tsuh "ffc ^ 
(a.d. 954-958) of the Hou Cou dynasty, among tribute-gifts sent from 
Kwa cou JR. m in Kan-su. 9 The Kirgiz received precious materials for 
the dress of their women from An-si (Parthia), Pei-t'ih 3b S (Bisbalik, 
in Turkistan) , and the Ta-si ;*C Jt (Tadzik, the Arabs). The Arabs made 
pieces of brocade of such size that the weight of each equalled that of 

twenty camel-loads. Accordingly these large pieces were cut up into 


1 Herodotus, vin, 120. 

2 Yates, Textrinum Antiquorum, pp. 366-368. 

* xxxm, 19, § 63. 

4 At the Court of the Persian kings there was a special atelier for the weaving 
of silken, gold, and silver fabrics, — styled siar baf xdne (E. Kaempfer, Amoenitatum 
exoticarum fasciculi V, p. 128, Lemgoviae, 1712). 

5 Lian Su, Ch. 54, p. 13 b. Hwa is the name under which the Ephthalites first 
appear in Chinese history (Chavannes, Documents sur les Tou-kiue occidentaux, 
p. 222). 

6 Kiu T'an Su, Ch. 198, p. 10 b (see also Lian $u, Ch. 54, p. 14 b; and Sui Su 
Ch. 83, p. 7 b). Huan Tsah refers to brocade in his account of Persia (Ta T'an si 
yil ki, Ch. 11, p. 17 b, ed. of Sou San ko ts'un Su). 

* Sui ht, hc^ & Hi # H J$ ^ Jft H in Lian Su, I. c. 

8 Cf. Loan-Words in Tibetan, No. 118. 

9 Wu tai U, Ch. 74, p. 3 b; Kiu Wu Tai H, Ch. 138, p. 1 b. 

Persian Textiles — Brocades 489 

twenty smaller ones, so that they could be accommodated on twenty 
camels, and were presented once in three years by the Arabs to the 
Kirgiz. The two nations had a treaty of mutual alliance, shared also 
by the Tibetans, and guaranteeing protection of their trade against the 
brigandage of the Uigur. 1 The term hu kin i$ ^ ("brocades of the Hu," 
that is, Iranians) is used in the Kwan yU ki Mf J8£ IS 2 with reference to 
Khotan. 3 The Iranian word for these textiles, though not recognized 
heretofore, is also recorded by the Chinese. This is & tie, anciently 
*d2iep, dziep, diep, dib, 4 being the equivalent of a Middle-Persian form 
*dib or *dep, 5 corresponding to the New-Persian word dibd ("silk bro- 
cade," a colored stuff in which warp and woof are both made of silk), 
dibah ("gold tissue"), Arabicised dibadZ ("vest of brocade, cloth of gold"). 
The fabric as well as the name come from Sasanian Persia, and were 
known to the Arabs at Mohammed's time. 6 The Chinese term occurs 
as a textile product of Persia in the Sui §u (Ch. 83, p. 7 b ). At a much 
earlier date it is cited in the Han Annals (Hou Han iw, Ch. 116, p. 8) 
as a product of the country of the Ai-lao in Yun-nan. This is not 
surprising in view of the fact that at that period Yun-nan, by way of 
India, was in communication with Ta Ts'in: in a.d. 120 Yuh Yu Tiao 
W. $3 M, King of the country T'an W, presented to the Chinese em- 
peror musicians and jugglers, who stated that "they had come from 
the Mediterranean M ©, which is the same as Ta Ts'in, and that 
south-west from the Kingdom of T'an there is communication with 
Ta Ts'in." The commentator of the Han Annals refers to the Wai kwo 
cwan $V 19 i$ 7 as saying that the women of Cu-po ft W (Java) make 
white tie and ornamented cloth ^£ ffi. The character JfJ po ("silk"), 
preceding the term tie in the Han Annals, represents a separate item, and 

1 T'an Sw, Ch. 217 B, p. 18; T'ai p'in hwan yil ki, Ch. 199, p. 14. Cf. Deveria, 
in Centenaire de l'Ecole des Langues Orientales, p. 308. 

; Ch. 24, p. 7 b. Regarding the various editions of this work, see p. 251. 

8 Likewise in the Sung Annals with reference to a tribute sent from Khotan 
in 961 (Chavannes and Pelliot, Traite" manich^en, p. 274). Regarding Persian 
brocades mentioned by mediaeval writers, see Francisque-Michel, Recherches sur 
le commerce, la fabrication et l'usage des 6toffes de soie d'or et d'argent, Vol. I, 
PP- 315-317. Vol. II, pp. 57-58 (Paris, 1852, 1854). 

4 According to the Yi ts'ie kin yin i (Ch. 19, p. 9 b), the pronunciation of the 
character tie was anciently identical with that of §|§ (see No. 70), and has the fan 
ts'ie tl $H; that is, t'iap, *diab, d'ab. The T'an iu Si yin (Ch. 23, p. I b) indicates 
the same fan ts'ie by means of ^ i$. The phonetic element J§ serves for the 
transcription of Sanskrit dvipa (Pelliot, Bull, de l'Ecole frangaise, Vol. IV, p. 357). 

5 A Pahlavi form depdk is indicated by West (Pahlavi Texts, Vol. I, p. 286); 
hence Armenian dipak. 

6 C. H. Becker, Encyclopaedia of Islam, Vol. I, p. 967. 

7 Cf. Journal asiatique, 1918, II, p. 24. 

490 Sino-Iranica 

is not part of the transcription, any more than the word t$ kin, which 
precedes it in the Sui Annals; but the combination of both po and kin 
with tie indicates and confirms very well that the latter was a brocaded 
silk. Hirth 1 joins po with tie into a compound in order to save the 
term for his pets the Turks. "The name po-tie is certainly borrowed 
from one of the Turki languages. The nearest equivalent seems to be 
the Jagatai Turki word for cotton, pakhta." There are two fundamental 
errors involved here. First, the Cantonese dialect, on which Hirth 
habitually falls back in attempting to restore the ancient phonetic 
condition of Chinese, does not in fact represent the ancient Chinese 
language, but is merely a modern dialect in a far-advanced stage of 
phonetic decadence. The sounds of ancient Chinese can be restored 
solely on the indications of the Chinese phonetic dictionaries and on the 
data of comparative Indo-Chinese philology. Even in Cantonese, 
po-tie is pronounced pak-tip, andvit is a prerequisite that the foreign 
prototype of this word terminates in a final labial. The ancient pho- 
netics of IfJ § is not pak-ta, but *bak-dzip or *dip, and this bears no 
relation to pakhta. Further, it is impossible to correlate a foreign 
word that appears in China in the Han period with that of a com- 
paratively recent Turkish dialect, especially as the Chinese data rela- 
tive to the term do not lead anywhere to the Turks; and, for the rest, 
the word pakhta is not Turkish, but Persian, in origin. 2 Whether the 
term tie has anything to do with cotton, as already stated by Cha- 
vannes, 3 is uncertain; but, in view of the description of the plant as 
given in the Nan & 4 or Lian iw, 5 it may be granted that the term po-tie 
was subsequently transferred to cotton. 

The ancient pronunciation of po-tie being *bak-dib, it would not be 
impossible that the element bak represents a reminiscence of Middle 
Persian pambak ("cotton"), New v Persian panpa (Ossetic bambag, 
Armenian bambak). This assumption being granted, the Chinese term 
po-tie { = Middle Persian *bak-dib = pambak dip) would mean "cotton 
brocade" or "cotton stuff." Again, po-tie was a product of Iranian 
regions: kin siu po tie dfe $8 6 & is named as a product of K'ah (Sog- 
diana) in the Sasanian era; 6 and, as has been shown, po-tie from Parthia 

1 Chao Ju-kua, p. 218. 

2 Steingass, Persian-English Dictionary, p. 237. 

1 Documents sur les Tou-kiue occidentaux, p. 352. 

4 Ch. 79, p. 6 b. 

5 Ch. 54, p. 13 b. Cf. Chavannes, ibid., p. 102; see also F. W. K. Muller, 
Uigurica, II, pp. 70, 105. 

• Sui lu, Ch. 83, p. 4. Hence *bak-dlb may also have been a Sogdian word. 

Persian textiles — Brocades 491 

is specially named. Po-tie, further, appears in India; 1 and as early as 
a.d. 430 Indian po-tie was sent to China from Ho-lo-tan MI ft "# on Java. 2 
According to a passage of the Kiu T'an iw, 3 the difference between ku- 
pei (Sanskrit karpasa)* and po-tie was this, that the former was a coarse, 

1 Nan ft, Ch. 78, p. 7 a. 

2 Sun $u, Ch. 97, p. 2 b. 

* Ch. 197, p. 1 b, indicated by Pelliot (Bull, de I'Ecole frangaise, Vol. Ill, 
p. 269). 

4 It is evident that the transcription ku-pei is not based directly on Sanskrit 
karpasa; but I do not believe with Watters (Essays on the Chinese Language, 
p. 440) and Hirth (Chau Ju-kua, p. 218) that Malayan kdpas is at the root of the 
Chinese form, which, aside from the lack of the final s, shows a peculiar vocalism that 
cannot be explained from Malayan. Of living languages, it is Bahnar kopaih ("cot- 
ton") which presents the nearest approach to Chinese ku-pei or ku-pai. It is there- 
fore my opinion that the Chinese received the word from a language of Indo-China. 
The history of cotton in China is much in need of a revision. The following case 
is apt to show what misunderstandings have occurred in treating this subject. 
Ku-lun (*ku-dzun, *ku-dun) "^ ^ is the designation of a cotton-like plant grown 
in the province of Kwei-cou ^£ M ; the yarn is dyed and made into pan pu i§£ ^ftj . 
This is contained in the Nan Yue li }$j ^ Jj»£ by Sen Hwai-yuan iffu f§[ j|£ of the 
fifth century (Pen ts'ao kan mu, Ch. 36, p. 24). Schott (Altaische Studien, III, 
Abh. Berl. Akad., 1867, pp. 137, 138; he merely refers to the source as "a descrip- 
tion of southern China," without citing its title and date), although recognizing that 
the question is of a local term, proposed, if it were permitted to read kutun instead 
of kutun, to regard the word as an indubitable reproduction of Arabic qu\un, which 
resulted in the coton, cotton, kattun, etc., of Europe. Mayers then gave a similar 
opinion; and Hirth (Chau Ju-kua, p. 219), clinging to a Fu-cbu pronunciation 
ku-tiin (also Waiters, Essays, p. 440, transcribes ku-tun), accepted the alleged 
derivation from the Arabic. This, of course, is erroneous, as in the fifth century 
there was no Arabic influence on China, nor did the Arabs themselves then know 
cotton. It would also be difficult to realize how a plant of Kwei-cou could have 
been baptized with an Arabic name at that or any later time. Moreover, ku-lun 
is not a general term for "cotton" in Chinese; the above work remains the only 
one in which it has thus far been indicated. Ku-lun, as Li 5i-£en points out, is a 
tree-cotton /fC £& (Bombax malabaricum) , which originated among the Southern 
Barbarians (Nan Fan fff ^), and which at the end of the Sung period was trans- 
planted into Kian-nan. It is very likely that, as stated by Stuart (Chinese Materia 
Medica, p. 197), the cotton-tree was known in China from very ancient times, and 
that its product was used in the manufacture of cloth before the introduction of the 
cotton-plant (Gossypium herbaceum). In fact, the same work Nan yiie li reports, 
"None of the Man tribes in the kingdom Nan-£ao rear silkworms, but they merely 
obtain the seeds of the so-lo (*sa-la) |£ ff| tree, the interior of which is white and 
contains a floss that can be wrought like silk and spun into cloth; it bears the name 
so-lo lun twan |£ H $1 Jg." The Fan yuli # g| J& of Cu Mu jffc f§ of the Sung 
period alludes to the same tree, which is said to be from thirty to fifty feet in height. 
The Ko ku yao lun (Ch. 8, p. 4 b; ed. of Si yin hiian ts'un Su) speaks of cotton stuffs 
!lE ft £& ( = ^S : tou-lo= Sanskrit tula) which come from the Southern Barbarians, 
Tibet (Si-fan), and Yun-nan, being woven from the cotton in the seeds of the so-lo 
tree, resembling velvet, five to six feet wide, good for making bedding and also clothes. 
The Tien hi writes the word ^ H (G. Souufe, Bull, de I'Ecole francaise, Vol. VIII, 
p. 343). Sa-la is the indigenous name of the tree; sa-la is still the Lo-lo designation 

492 Sino-Iranica 

and the latter a fine textile. In the Glossary of the T'ang Annals the 
word tie is explained as "fine hair" $B 3s and "hair cloth" ^ ^ftf; these 
terms indeed refer to cotton stuffs, but simultaneously hint at the fact 
that the real nature of cotton was not yet generally known to the Chinese 
of the T'ang period. In the Kwan yii ki, po-tie is named as a product of 
Turf an; the threads, it is said, are derived from wild silkworms, and 
resemble fine hemp. 

Russian altabds ("gold or silver brocade," "Persian brocade": 
Dal'), Polish altembas, and French altobas, in my opinion, are nothing 
but reproductions of Arabic-Persian al-dibadz", discussed above. The 
explanation from Italian alto-basso is a jocular popular etymology; and 
the derivation from Turkish altun ("gold") and b'az ("textile") 1 is 
likewise a failure. The fact that textiles of this description were subse- 
quently manufactured in Europe has nothing to do, nor does it conflict, 
with the derivation of the name which Inostrantsev wrongly seeks in 
Europe. 2 In the seventeenth century the Russians received altabds 
from the Greeks; and Ibn Rosteh, who wrote about a.d. 903, speaks 
then of Greek dlbadz? According to Makkari, dibadZ were manufac- 
tured by the Arabs in Almeria, Spain, 4 the centre of the Arabic silk 
industry. 5 

70. Hlf£f fa-ten, *dap ( = lil) 6 -dan ( = ^£), tap-tan, woollen rugs. 
The name of this textile occurs in the Wei lio of the third century a.d. 
as a product of the anterior Orient (Ta Ts'in) , 7 and in the Han Annals 

for cotton (Vial, Dictionnaire francais lo-lo, p. 97). Likewise it is sa-la in P'u-p'a, 
so-1'6 in Co-ko {Bull, de I'Ecole frangaise, Vol. IX, p. ,554). In the same manner I 
believe that *ku-dzun was the name of the same or a similar tree in the language of 
the aborigines of Kwei-2ou. Compare Lepcha ka-cuk ki kun ("cotton-tree"), Sih-p'o 
ga-dun ("cotton-tree"), given by J. F. Needham (Outline Grammar of the Singpho 
Language, p. 90, Shillong, 1889), and MeoV^oa ("cotton"), indicated by M. L. 
Pierlot (Vocabulaire m£o, Actes du XIV 8 Congres int. des Orientalistes Alger 
1905, pt. I, p. 150). 

1 Proposed by Savel'ev in Erman's Archiv, Vol. VII, 1848, p. 228. 

1 K. Inostrantsev, Iz istorii starinnyx tkanei {Zapiski Oriental Section Russian 
Archaeol. Soc, Vol. XIII, 1901, pp. 081-084). 

1 G. Jacob, Handelsartikel, p. 7; Waren beim arabisch-nordischen Verkehr, 
p. 16. 

' G. Migeon, Manuel d'art musulman, Vol. II, p. 420. 

8 Defremerv, Journal asiatique, 1854, p. 168; Francisque-Michel, Recherches 
sur le commerce, la fabrication et l'usage des 6toffes de soie, d'or et d'argent, Vol. I, 
pp. 232, 284-290 (Paris, 1852). 

6 The/a» ts'ie is $E #$f; that is, *du-kiap = d'iap (Yi ts'ie kinyin i, Ch. 19, p. 9 b), 
or *£ Wi *du-hap = dap (Hou Han $u, Ch. 118, p. 5 b). 

7 F. Hirth, China and the Roman Orient, pp. 71, 112, 113, 255. T'a-ten of five 
and nine colors are specified. 

Persian Textiles — Rugs 493 

as a product of India. 1 In the Sui Annals it appears as a product of 
Persia. 2 Chavannes has justly rejected the fantastic explanation given 
in the dictionary Si min, which merely rests on an attempt at punning. 
The term, in fact, represents a transcription that corresponds to a 
Middle-Persian word connected with the root Vtab ("to spin"): 
cf. Persian tdftan ("to twist, to spin"), tabab ("he spins"), tdfta or tafte 
("garment woven of linen, kind of silken cloth, taffeta"). Greek Tcnrr/s 
and rawriTlov (frequent in the Papyri; rairLdvcfroi, " rug- weavers ") are 
derived from Iranian. 3 There is a later Attic form Sd-n-ts. The Middle- 
Persian form on which the Chinese transcription is based was perhaps 
*taptah, tapetan, -an being the termination of the plural. The Persian 
word resulted in our taffeta (med. Latin taffata, Italian taffeta, Spanish 

71. To the same type as the preceding one belongs another Chinese 
transcription, $H if io{Vo)-pi, $H %$ tso-pH, or #1 & tso-pi, dance- 
rugs sent to China in a.d. 718 and 719 from Maimargh and Bukhara 
respectively. 4 These forms correspond to an ancient *ta-bik (1st or %$) 
or *ta-bi5 (4$), and apparently go back to two Middle-Persian forms 
*tabix and *tabe5 or *tabi5 (or possibly with medial p)} 

72. More particularly we hear in the relations of China with 
Persia about a class of textiles styled yiie no pu MW ^ft». 6 As far as I 
know, this term occurs for the first time in the Annals of the Sui Dy- 
nasty (a.d. 590-617), in the notice on Po-se (Persia). 7 This indicates 
that the object in question, and the term denoting it, hailed from Sasa- 
nian Persia. 

1 E. Chavannes, Les Pays d'occident d'apres le Heou Han Chou {T'oung Pao, 
1907, p. 193). Likewise jin the Nan Si (Ch. 78, p. 5 b) and in Cao Zu-kwa (trans- 
lation of Hirth and Rockhill, p. in). 

2 Sui I«, Ch. 83, p. 7 b. 

s P. Horn, Grundriss iran. Phil., Vol. I, pt. 2, p. 137. Noldeke's notion 
(Persische Studien, II, p. 40) that Persian tanbasa ("rug, carpet") should be derived 
from the Greek word, in my opinion, is erroneous. 

* Chavannes, T'oung Pao, 1904, p. 34. 

5 These two parallels possibly are apt to shed light on the Old High-German 
duplicates teppih and teppid. The latter has been traced directly to Italian tappeto 
(Latin tapete, tapetum), but the origin of the spirant x hi teppih has not yet been 
explained, and can hardly be derived from the final /. Would derivation from an 
Iranian source, direct or indirect, be possible? 

8 According to Hirth (Chau Ju-kua, p. 220), "a light cotton gauze or muslin, 
of two kinds, pure white, and spangled with gold"; but this is a doubtful explana- 

7 Sttt Su, Ch. 83, p. 7 b. This first citation of the term has escaped all previous 
writers on the subject, — Hirth, Chavannes, and Pelliot. From the Sui Su the text 
passed into the T'ai p'in hwan yil ki (Ch. 185, p. 18 b). 

494 Sino-Iranica 

In the T'ang Annals we read that in the beginning of the period 
K'ai-yuan (a.d. 713-741) the country of K'ah (Sogdiana), an Iranian 
region, sent as tribute to the Chinese Court coats-of-mail, cups of rock- 
crystal, bottles of agate, ostrich-eggs, textiles styled yiie no, dwarfs, 
and dancing-girls of Hu-siian $] W. (Xwarism). 1 In the Ts'efu yuan kwei 
the date of this event is more accurately fixed in the year 718. 2 The 
Man $u, written by Fan Co of the T'ang period, about a.d. 860, 3 men- 
tions yiie no as a product of the Small P'o-lo-men /b ^ ft PI (Brah- 
mana) country, which was conterminous with P'iao HI (Burma) and 
Mi-6'en (*Midzen) ffl E5. 4 This case offers a parallel to the presence 
of tie in the Ai-lao country in Yiin-nan. 

The Annals of the Sung mention yiie no as exported by the Arabs 
into China. 5 The Lin wai tai ta, e written by Cou K'u-fei in 11 78, men- 
tions white yiie-no stuffs in the countries of the Arabs, in Bagdad, and 
yiie-no stuffs in the country Mi fH. 

Hirth 7 was the first to reveal the term yiie no in Cao Zu-kwa, who 
attributes white stuffs of this name to Bagdad. His transcription, yiit- 
nok, made on the basis of Cantonese, has no value for the phonetic 
restoration of the name, and his hypothetical identification with cut- 
tanee must be rejected; but as to his collocation of the second element 
with Marco Polo's nac, he was on the right trail. He was embarrassed, 
however, by the first element yiie, "which can in no way be explained 
from Chinese and yet forms part of the foreign term." Hence in his 
complete translation of the work 8 he admits that the term cannot as 
yet be identified. His further statement, that in the passage of the 
T'an iw, quoted above, the question is possibly of a country yiie-no 
(Bukhara), rests on a misunderstanding of the text, which speaks only 
of a textile or textiles. The previous failures in explaining the term 
simply result from the fact that no serious attempt was made to restore 

1 Cf. Chavannes, Documents sur les Tou-kiue occidentaux, pp. 136, 378, 
with the rectification of Pelliot {Bull, de I'Ecole frangaise, Vol. IV, 1904, p. 483). 
Regarding the dances of Hu-siian, see Kin H hwiyiian kiao k'an ki j£ if^ "H" Jfc $t 
Wl 12 (p. 3), Critical Annotations on the Kin si hwi yuan by Li Sah-kiao ^£ h ^ 
of the Sung (in Ki fu ts*un §u, t'ao 10). 

2 Chavannes, T'oung Pao, 1904, p. 35. 

3 See above, p. 468. 

* Man $u, p. 44 b (ed. of Yiin-nan pei Sen ci). Regarding Mi-5'en, see Pelliot, 
Bull, de I'Ecole frangaise, Vol. IV, p. 171. 

6 Sun U, Ch. 490; and Bretschneider, Knowledge possessed by the Chinese 
of the Arabs, p. 12. Bretschneider admitted that this product was unknown to him. 

6 Ch. 3, pp. 2-3. 

7 Lander des Islam, p. 42 (Leiden, 1894). 

8 Chau Ju-kua, p. 220. 

Persian Textiles — Yue No 495 

it to its ancient phonetic condition. 1 Moreover, it was not recognized 
that yiie no represents a combination of two Iranian words, and that 
each of these elements denotes a particular Iranian textile. 

(1) The ancient articulation of what is now sounded yiie ^ was 
*vat, va5, wia5, or, with liquid final, *var or *val. 2 Thus it may well 
be inferred that the Chinese transcription answers to a Middle-Persian 
form of a type *var or *val. There is a Persian word barnu or barnun 
("brocade")? void, which means "a kind of silken stuff," 3 and balds, 
"a kind of fine, soft, thin armosin silk, an old piece of cloth, a kind of 
coarse woollen stuff." 4 

(2) Hr no corresponds to an ancient *nak, 5 and is easily identified 
with Persian nax (nakh), "a carpet beautiful on both sides, having a 
long pile; a small carpet with a short pile; a raw thread of yarn of any 
sort," 6 but also "brocade." The early mention of the Chinese term, 
especially in the Sui Annals, renders it quite certain that the word nak 
or nax was even an element of the Middle-Persian language. Hither- 
to it had been revealed only in mediaeval authors, the Yuan fao pi H t 

1 De Goeje's identification of yue-no pu with djannabi (in Hirth, Lander des Islam, 
p. 61) is a complete failure: pu ("cloth") does not form part of the transcription, 
which can only be read vaS-nak, var-nak, or val-nak. Tsuboi Kumazo (Actes XII" 
Congres international des Orientalistes Rome 1899, Vol. II, p. 112) has already 
opposed this unfortunate suggestion. 

2 For examples, see Chavannes, Memoires historiques de Se-ma Ts'ien, 
Vol. IV, p. 559; and particularly cf. Pelliot, Journal asiatique, 1914, II, p. 392. 

1 Steingass, Persian-English Dictionary, p. 1453. Horn (Grundriss iran. 
Phil., Vol. I, pt. 2, p. 29) translates the word "a fine stuff, " and regards it as a loan- 
word from Greek ftijXov ("veil"), first proposed, I believe, by Noldeke (Persische 
Studien, II, p. 39). This etymology is not convincing to me. On the contrary, 
vala is a genuine Persian word, meaning "eminent, exalted, high, respectable, sub- 
lime, noble"; and it is quite plausible that this attribute was transferred to a fine 
textile. It was, further, the Persians who taught the Greeks lessons in textile art, 
but not the reverse. F. Justi (Iranisches Namenbuch, p. 516) attributes to vala 
also the meaning "banner of silk." 

4 Steingass, op. cit., p. 150. The Iranian character of this word is indicated 
by WaxI palds, Sariqoll palus ("coarse woollen cloth") of the Pamir languages. 
Perhaps also Persian bat ("stuff of fine wool"), WaxI bot, Sariqoll bei (cf. W. Toma- 
schek, Pamirdialekte, Sitzber. Wiener Akad., 1880, p. 807) may be enlisted as possible 
prototypes of Chinese *vat, val; but I do not believe with Tomaschek that this 
series bears any relation to Sanskrit pa\\a and lafa or Armenian lotik ("mantle"). 
The latter, in my opinion, is a loan-word from Greek X<£5i£ ("cover, rug"), that 
appears in the Periplus (§ 24) and in the Greek Papyri of the first century a.d. 
(T. Reil, Beitrage zur Kenntnis des Gewerbes im hellenistischen ^Egypten, p. 118). 

s See, for instance, T'oung Pao, 1914, p. 77, and 1915, p. 8, where the character 
in question serves for transcribing Tibetan nag. It further corresponds to nak 
in Annamese, Korean, and Japanese, as well as in the transcriptions of Sanskrit 

8 Steingass, Persian-English Dictionary, p. 1391. 

496 Sino-Iranica 

Yuan H, Ibn Batata, Rubruk, Marco Polo, Pegoletti, etc. 1 W. Bang 
has shown in a very interesting essay 2 that also the Codex Cumanicus 
contains the term nac (Cumanian), parallel with Persian nagh and Latin 
nachus, in the sense of "gold brocades," and that the introitus natorum 
et nascitorum of the books of tax-rates of Genoa about 1420 refers to 
these textiles, and has nothing to do with the endowment of the new- 
born, as had been translated. Bang points out also "nachi, a kinde 
of slight silke wouen stuff e" in Florio, "Queen Anna's New World of 
Words" (London, 161 1). In mediaeval literature the term nac, nak, 
naque, or nachiz occurs as early as the eleventh century, and figures in 
an inventory of the Cathedral of Canterbury of the year 13 15. 

73. 1$ M hu-na, *yu-na, a textile product of Persia 3 (or IS -f*$). 4 An 
ancient Iranian equivalent is not known to me, but must be supposed 
to have been *7una or *guna. This word may be related to Sighnan 
(Pamir language) ghdun ("coarse sack"), Kashmir gun, Sanskrit gonl; b 
Anglo-Indian gunny, gunny-bag, trading-name of the coarse sacking 
and sacks made from the fibre of the jute. 6 

74. W. fan, *dan, *tan, a textile product of Persia, likewise men- 
tioned in the Sui Annals. This is doubtless the Middle-Persian des- 
ignation of a textile connected with the root Vtan ("to spin"), of 
which several Middle-Persian forms are preserved. 7 Compare Avestan 
tanva, Middle Persian tanand, Persian taniban, tanando ("spider"), 
and, further, Persian tan-basa, tan-blsa ("small carpet, rug"); tanid 
("a web"); tantdan ("to twist, weave, spin"). 

75. Jb "n $!) sa-ha-la or J§ &« p §l s so-ha-la, of green color, is men- 

1 See E. Bretschneider, Notices of the Mediaeval Geography, p. 288, or Me- 
diaeval Researches, Vol. II, p. 124; Yule, Cathay, new ed. by Cordier, Vol. Ill, 
PP- I55 -I 56, 169; Yule, Marco Polo, Vol. I, pp. 63, 65, 285; W. Heyd, Histoire 
du commerce du levant au moyen age, p. 698; and, above all, P.-Michel, Recherches 
sur le commerce etc., des £toffes de soie, Vol. I, pp. 261-264. A. Houtum-Schindler 
(Journal As. Soc. Bengal, Vol. VI, 1910, p. 265) states that nax occurs in a letter of 

2 Ueber den angeblichen "Introitus natorum et nascitorum" in den Genueser 
Steuerbuchern, in Bull, de la Classe des Lettres de V Academie royale de Belgiaue, 
No. 1, 1912, pp. 27-32. 

* Sui lu, Ch. 83, p. 7 b. / 

4 T'ai p'in hwan yii ki, Ch. 185, p. 18 b. 

5 W. Tomaschek, Pamirdialekte (Sitzber. Wiener Akad., 1880, p. 808). 

6 Yule, Hobson-Jobson, p. 403. 

7 Salemann, Grundriss iran. Phil., Vol. I, pt. 1, p. 303. 

8 This transcription is given in the C'an wu li U $$/} ]^i by Wen Cen-heh $£ 
JH -^ of the Ming (Ch. 8, p. 1 b; ed. of Yue ya Van Is'un Su). He describes the 
material as resembling sheep-wool, as thick as felt, coming from the Western 
Regions, and very expensive. 

Persian Textiles — Woollen Stuffs 497 

tioned in the Ming history as having been sent as a present in 1392 from 
Samarkand. The Ming Geography, as stated by Bretschneider, 1 
mentions this stuff as a manufacture of Bengal and So-li, saying that 
it is woven from wool and is downy. There is a red and a green kind. 
Bretschneider's view, that by sa-ha-la the Persian ia/ is intended, must 
be rejected. 2 In the Yin yai leh Ian of 1416, sa-ha-la is enumerated 
among the goods shipped from Malacca, being identified by Groene- 
veldt with Malayan saklat or sahalat. 3 Sa-ha-la is further mentioned 
for Ormuz and Aden. 4 

In the Ko ku yao lun $&• "& 5* H, written by Ts'ao Cao W B3 in 
1387, revised and enlarged in 1459 by Wan Tso 3E t*c, 5 we meet this 
word in the transcription $§ ( = $5) M ffl sa-hai-la, 6 which is said to 
come from Tibet IS # in pieces three feet in width, woven from wool, 
strong and thick like felt, and highly esteemed by Tibetans. Under the 
heading p'u-lo laf ft ( = Tibetan p'rug) 1 it is said in the same work that 
this Tibetan woollen stuff resembles sa-Jtai-la. 

Persian sakirlat, sagirlat, has been placed on a par with Chinese 
sa-ha-la by T. Watters 8 and A. Houtum-Schindler; 9 it is not this 
Persian word, however, that is at the root of Chinese sa-ha-la, but 
saqalat or saqalldt, also saqalat, saqalldf ("scarlet cloth"). Dr. E. D. 
Ross 10 has been so fortunate as to discover in a Chinese-Persian vocabu- 
lary of 1549 the equation: Chinese sa-ha-la = Persian saqalat. This settles 
the problem definitely. There is, further, Persian saqldfiin or saqldfin, 
said to mean "a city in Rum where scarlet cloth is made, scarlet cloth 
or dress made from it." The latter name is mentioned as early as 
a.d. 1040 and 1 150 by Baihaki and Edrlsi respectively. 11 According to 
Edrisi, it was a silk product of Almeria in Spain, which is doubtless 
meant by the city of Rum. Yaqut tells of its manufacture in Tabriz, 

1 Mediaeval Researches, Vol. II, p. 258. 

a Regarding the Chinese transcription of this Persian word, see Rockhill, T'oung 
Pao, 191 5, p. 459. 

3 Notes on the Malay Archipelago, p. 253. 

* Rockhill, T'oung Pao, 1915, pp. 444, 606, 608. It does not follow from the 
text, however, that sa-ha-la was a kind of thin veiling or gauze, as the following 
term (or terms) £|| -{5£ is apparently a matter in itself. 

5 Ch. 8, p. 4 b (ed. of Si yin huan Is'un Su). 

9 This mode of writing is also given in the C'an wu li, cited above. 

7 T'oung Pao, 1914, p. 91. 

8 Essays on the Chinese Language, p. 342. v 

9 Journal As. Soc. Bengal, Vol. VI, 1910, p. 265. 

10 Journal As. Soc. Bengal, Vol. IV, 1908, p. 403. 

11 Yule, Hobson-Jobson, p. 861. 

498 Sino-Iranica 

so that the Chinese reference to Samarkand becomes intelligible. The 
Chinese reports of sa-ha-la in India, Ormuz, and Aden, however, evi- 
dently refer to European broadcloth, as does also Tibetan sag-lad. 1 

The Ain-i Akbari speaks of sukldt (saqaldt) of Rum (Turkey), 
FarangI (Europe), and Purtagall (Portugal); and the Persian word is 
now applied to certain woollen stuffs, and particularly to European 

The Persian words sakirlat and saqaldt are not interrelated, as is 
shown by two sets of European terms which are traced to the two 
Persian types: sakirlat is regarded as the ancestor of "scarlet" (med. 
Latin scarlatum, scarlata; Old French escarlate, New French ecarlate, 
Middle English scarlat, etc.); saqldtun or siqldtun is made responsible 
for Old French siglaton, Provencal sisclaton (twelfth century), English 
obs. ciclatoun (as early as 1225), Middle High German cicldt or sigldt. 
Whether the alleged derivations from the Persian are correct is a de- 
batable point, which cannot be discussed here; the derivation of siglaton 
from Greek Ku/cXds {cyclas), due to Du Cange, is still less plausible. 2 
Dr. Ross {I.e.) holds that "the origin of the word scarlet seems to be 
wrapped in mystery, and there seems to be little in favor of the argu- 
ment that the word can be traced to Arabic or Persian sources." 

76. Toward the close of the reign of Kao Tsun ftS Hx, better known 
as Wen C'eh j£ $ (a.d. 452-465) of the Hou Wei dynasty (386-532), 
the king of Su-le (Kashgar) sent an emissary to present a garment 
(kdsdya) of C akyamuni Buddha, over twenty feet in length. On ex- 
amination, Kao Tsuh satisfied himself that it was a Buddha robe. It 
proved a miracle, for, in order to get at the real facts, the Emperor 
had the cloth put to a test and exposed to a violent fire for a full day, but 
it was not consumed by the flames. All spectators were startled and 
spell-bound. 3 This test has repeatedly^ been made everywhere with 
asbestine cloth, of which many examples are given in my article 
"Asbestos and Salamander." 4 The Chinese themselves have recog- 
nized without difficulty that this Buddha relic of Kashgar was made 
of an asbestine material. In the Lu Van kun H k%? a modern work, 

1 See Loan-Words in Tibetan, No. 119. 

2 Cf. also F.-Michel, Recherches sur le commerce etc., des 6toffes de soie, 
Vol. I, pp. 233-235. The Greek word in question does not refer to a stuff, but to a 
robe (kvkX&s, "round, circular," scil., e<rdr)s, "a woman's garment with a border all 
round it "). Cycladatus in Suetonius (Caligula, lii) denotes a tunic with a rich border. 

3 Wei Su, Ch. 102, p. 4 b. 

4 T'oung Pao, 1915, pp. 299-373. 

5 Ed. of Ts'in lao Van ts'un $u, p. 40 (see above, p. 346). On p. 41 b there is a 
notice of fire-proof cloth, consisting of quotations from earlier works, which are 
all contained in my article. 

Persian Textiles — Asbestos 499 

which contains a great number of valuable annotations on subject- 
matters mentioned in the Annals, the ka$aya of Kashgar is identified 
with the fire-proof cloth of the Western Regions and Fu-nan (Camboja) ; 
that is, asbestos. 

During the K'ai-yuan and T'ien-pao periods (a.d. 713-755), Persia 
sent ten embassies to China, offering among other things "embroideries 
of fire-hair" {hwo mao siu iK ^ I&). 1 Chavannes 2 translates this term 
"des broderies en laine couleur de feu." In my opinion, asbestos is 
here in question. Thus the term was already conceived by Abel- 
Remtjsat. 3 I have shown that asbestos was well known to the Persians 
and Arabs, and that the mineral came from Badax§an. 4 An additional 

1 Van su, Ch. 221 b, p. 7. In the T'an hui yao (Ch. ioo, p. 4) this event is 
fixed in the year 750. 

2 Documents sur les Tou-kiue, p. 173. 

* Nouveaux melanges asiatiques, Vol. I, p. 253. The term hwo pu jK^fi ("fire- 
cloth") for asbestos appears in the Sun Su (Ch. 97, p. 10). The Chinese notions of 
textiles made from an "ice silkworm," possibly connected with Persia (cf. H. Mas- 
pero, Bull, de I'Ecole frangaise, Vol. XV, No. 4, 1915, p. 46), in my opinion, must 
be dissociated from asbestos; the Chinese sources (chiefly Wei lio, Ch. 10, p. 2 b) 
say nothing to the effect that this textile was of the nature of asbestos. Maspero's 
argumentation (ibid., pp. 43-45) in regard to the alleged asbestos from tree-bark, 
which according to him should be a real asbestine stuff, appears to me erroneous. 
He thinks that I have been misled by an inexact translation of S. W. Williams. 
First, this translation is not by Williams, but, as expressly stated by me (/. c, 
p. 372), the question is of a French article of d'Hervey-St.-Denys, translated into 
English by Williams. If an error there is (the case is trivial enough), it is not due to 
Williams or myself, but solely to the French translator, who merits Maspero's criticism. 
Second, Maspero is entirely mistaken in arguing that this translation should have 
influenced my interpretation of the text on p. 338. This is out of the question, as all 
this was written without knowledge of the article of St.-Denys and Williams, which 
became accessible to me only after the completion and printing of the manuscript, 
and was therefore relegated to the Addenda inserted in the proofs. Maspero's in- 
terpretation leads to no tangible result, in fact, to nothing, as is plainly manifest 
from his conclusion that one sort of asbestos should have been a textile, the other a 
kind of felt. There is indeed no asbestos felt. How Maspero can deny that Malayan 
bark-cloth underlies the Chinese traditions under notice, which refer to Malayan 
regions, is not intelligible to me. Nothing can be plainer than the text of the 
Liang Annals: "On Volcano Island there are trees which grow in the fire. The 
people in the vicinity of the island peel off the bark, and spin and weave it into cloth 
hardly a few feet in length. This they work into kerchiefs, which do not differ in 
appearance from textiles made of palm and hemp fibres," etc. (pp. 346, 347). What 
else is this but bark-cloth? And how could we assume a Malayan asbestine cloth 
if asbestos has never been found and wrought anywhere in the Archipelago? I 
trust that M. Maspero, for whose scholarship I have profound respect, will pardon 
me for not accepting his opinion in this case, and for adhering to my own inter- 
pretation. I may add here a curious notice from J. A. de Mandelslo's Voyages 
into the East Indies (p. 133, London, 1669): "In the Moluccaes there is a certain 
wood, which, laid in the fire, burns, sparkles, and flames, yet consumes not, "nd 
yet a man may rub it to powder betwixt his fingers." 

* T'oung Pao, 1915, pp. 327-328. 

500 Sino-Iranica 

text to this effect may be noted here. Ibn al-Faqih, who wrote in 
a.d. 902, has this account: "In Kirman there is wood that is not burnt 
by fire, but comes out undamaged. 1 A Christian 2 wanted to commit 
frauds with such wood by asserting that it was derived from the cross of 
the Messiah. Christian folks were thus almost led into temptation. A 
theologian, noting this man, brought them a piece of wood from Kir- 
man, which was still more impervious to fire than his cross-wood." 
According to P. Schwarz, 3 to whom we owe the translation of this 
passage, the question here is of fossilized forests. Most assuredly, how- 
ever, asbestos is understood. The above text of the Wei lu is thus by 
far the earliest allusion to asbestos from an Iranian region. 

The following notes may serve as additional information to my 
former contribution. Cou Mi M $? (1230-13 20), in his Ci ya Van tsa 
c'ao Je£ #i ^ H ^, mentions asbestine stuffs twice. 4 In one passage 
he relates that in his house there was a piece of fire-proof cloth (hwo 
hwan pu) over a foot long, which his maternal grandfather had once 
obtained in Ts'uan Cou ^. 'M (Fu-kien Province). 5 Visitors to his house 
were entertained by the experiment of placing it on the fire of a brazier. 
Subsequently Cao Mon-i fit j£l HI borrowed it from him, but never 
returned it. In the other text he quotes a certain Ho Ts'ih-fu W. W ^ 
to the effect that fire-proof cloth is said to represent the fibres of the 
mineral coal of northern China, burnt and woven, but not the hair of 
the fire-rodent (salamander). This is accompanied by the comment 
that coal cannot be wrought into fibres, but that now pu-hwei-mu 
-T J^ ?fc (a kind of asbestos) is found in Pao-tih (Ci-li). 6 A brief notice 
of asbestos is inserted in the Ko ku yao lun, 7 where merely the old fables 
are reiterated. Information on the asbestos of Ci-li Province will be 

1 Qazwlnl adds to this passage, "even if deft in fire for several days." 

a Qazwlnl speaks in general of charlatans. 

3 Iran im Mittelalter, p. 214. 

1 Ch. A, p. 20 b; and Ch. B, p. 25 b (ed. of Yue ya Van ts'un Su). 

5 This locality renders it almost certain that this specimen belonged to those 
imported by the Arabs into China during the middle ages (p. 331 of my article). 
The asbestos of Mosul is already mentioned in the Lin wai tai ta (Ch. 3, p. 4). 

8 The term pu-hwei-mu ("wood burning without ashes, incombustible wood") 
appears as early as the Sung period in the Cen lei pen ts'ao (Ch. 5, p. 35): it comes 
from San- tan (south-east portion of San-si and part of Ho-nan), and is now found 
in the Tse-lu mountains : \$fc |JL| . It is a kind of stone, of green and white color, 
looking like rotten wood, and cannot be consumed by fire. Some call it the root of 

7 Ch. 8, p. 4 (ed. of Si yin Man ts'un $u). In Ch. 7, p. 17, there is a notice on 
pu-hwei-mu stone, stated to be a product of Tse-cbu and Lu-han in San-si, and em- 
ployed for lamps. 

Persian Textiles — Asbestos 501 

found in the Kifu Vuh ci, 1 on asbestos of Se-SVan in the Se Pwan fun U* 
In the eighteenth century the Chinese noticed asbestos among the 
Portuguese of Macao, but the article was rarely to be found in the 
market. 3 HanzO Murakami discusses asbestos (^ $$, "stone cotton") 
as occurring in the proximity of Kin-cou 4£ ilfti in Sen-kin, Manchuria. 4 

In regard to the salamander, Francisque-Michel 5 refers to "Tradi- 
tions teratologiques de Berger de Xivrey" (Paris, Imprimerie royale, 
1836, pp. 457, 458, 460, 463) and to an article of Duchalais entitled 
"L'Apollon sauroctone" {Revue archiologique, Vol. VI, 1850, pp. 87-90); 
further to Mahudel in Memoires de litterature tiris des registres de 
V Acad&mie royale des inscriptions et belles-lettres, Vol. IV, pp. 634-647. 
Quoting several examples of salamander stuff from mediaeval romances, 
Francisque-Michel remarks, "Ces £toffes en poil de salamandre, qui 
vraisemblablement 6taient pass£es des fables des marchands dans celles 
des poetes, venaient de loin, comme ceux qui avaient par la beau jeu 
pour mentir. On en faisait aussi des manteaux; du moins celui de 
dame Jafite, du Roman de Gui le Gallois, en £tait." 

No one interested in this subject should fail to read chapter LII of 
book III of Rabelais' Le Gargantua et Le Pantagruel, entitled "Comment 
doibt estre prepare* et mis en ceuvre le celebre Pantagruelion." 

77. The word "drugget," spelled also droggitt, drogatt, druggit (Old French 
droguet, Spanish droguete, Italian droghetto) is thus denned in the new Oxford English 
Dictionary: "Ulterior origin unknown. Littr6 suggests derivation from drogue 
drug as 'a stuff of little value'; some English writers have assumed a derivation 
from Drogheda in Ireland, but this is mere wanton conjecture, without any histor- 
ical basis. Formerly kind of stuff, all of wool, or mixed of wool and silk or wool and 
linen, used for wearing apparel. Now, a coarse woollen stuff for floor-coverings, 
table-cloths, etc." The Century Dictionary says, "There is nothing to show a con- 
nection with drug." 

Our lexicographers have overlooked the fact that the same word occurs also 
in Slavic. F. Miklosich 8 has indicated a Serbian doroc ("pallii genus") and Magyar 
darocz ("a kind of coarse cloth"), but neglected to refer to the well-known Russian 
word dorogi or ddrogi, which apparently represents the source of the West-European 
term. The latter has been dealt with by K. Inostrantsev 7 in a very interesting 

1 Ch. 74, pp. 10 b, 13. 
1 Ch. 74, p. 25. 

3 Ao-men li Ho, Ch. B, p. 41. 

4 Journal Geol. Soc. Tokyo, Vol. XXIII, No. 276, 1916, pp. 333-336. The 
same journal, Vol. XXV, No. 294, March, 1918, contains an article on asbestos in 
Japan and Korea by K. Okada. 

5 Recherches sur le commerce, la fabrication et l'usage des eloffes de soie, d'or 
et d'argent, Vol. II, pp. 90, 462 (Paris, 1854). 

• Fremdworter in den slavischen Sprachen, Denk. Wiener Akad., Vol. XV, 
1867, p. 84. 

7 Iz istorii starinnix tkanei, Zapiski of the Russian Arch. Soc, Vol. XIII, 1902, 
p. 084. 

502 Sino-Iranica 

study on the history of some ancient textiles. According to this author, the dorSgi 
of the Russians were striped silken fabrics, which came from Gilan, Kasan, Kizylba§, 
Tur, and Yas in Persia. Dal' says in his Russian Dictionary that this silk was some- 
times interwoven with gold and silver. In 1844 Veltman proposed the identity of 
Russian dorogi with the Anglo-French term. Berezin derived it from Persian 
daradza ("kaftan"), which is rejected, and justly so, by Inostrantsev. On his part, 
he connects the word with Persian darai ("a red silken stuff"), 1 and invokes a 
passage in Veselovski's "Monuments of Diplomatic and Commercial Relations of 
Moscovite Rus with Persia," in which the Persian word darai is translated by 
Russian dorogi. This work is unfortunately not accessible to me, so I cannot judge 
the merits of the translation; but the mere fact of rendering dorogi by darai would 
not yet prove the actual derivation of the former from the latter. For philological 
reasons this theory seems to me improbable: it is difficult to realize that the Russians 
should have made dorogi out of a Persian darai. All European languages have con- 
sistently preserved the medial g, and this cannot be explained from darai. 
Another prototype therefore, it seems to me, comes into question; and this probably 
is Uigur torgu, Jagatai torka, Koibal torga, Mongol torga(n), all with the meaning 
"silk." 2 It remains to search for the Turkish dialect which actually transmitted 
the word to Slavic. 

1 Mentioned, for instance, in the list of silks in the Ain-i Akbari (Blochmann's 
translation, Vol. I, p. 94). 

3 Cf. T'oung Pao, 1916, p. 489. 


78. V$ i& hu-lo, *xu-lak, perhaps also *fu-lak, *fu-rak, a product of 
Persia, 1 which is unexplained. In my opinion, this word may cor- 
respond to a Middle Persian *furak = New Persian biirak, bur a, Arme- 
nian porag ("borax"). Although I am not positive about this identifica- 
tion, I hope that the following notes on borax will be welcome. It is 
well known that Persia and Tibet are the two great centres supplying 
the world-market with borax. The ancient Chinese were familiar with 
this fact, for in the article on Po-se (Persia) the T'ai p'in hwan yii ki 1 
states that "the soil has salty lakes, which serve the people as a substi- 
tute for salt " (H W N *& A ft ■ Bfc) . Our own word "borax " (the x is 
due to Spanish, now written borraj) comes from Persian, having been 
introduced into the Romanic languages about the ninth century by 
the Arabs. Russian burd was directly transmitted from Persia. Like- 
wise our "tincal, tincar" (a crude borax found in lake-deposits of 
Persia and Tibet) is derived from Persian tinkdr, tankdl, 3 or tangar, 
Sanskritized fankana, {anka, fanga, fagara;* Malayan tingkal; Kirgiz 
ddndkdr, Osmanli tdngar} Another Persian word that belongs to this 
category, $ora ("nitre, saltpetre"), has been adopted by the Tibetans 
in the same form $o-ra, although they possess also designations of their 
own, ze-ts'wa, ba-ts'wa ("cow's salt"), and is'a-la. The Persian word is 
Sanskritized into sordka, used in India for nitre, saltpetre, or potassium 
nitrate. 6 

79. The relation of Chinese nao-la ("sal ammoniac, chloride of 
sodium") 7 to Persian nulddir or naulddir is rather perspicuous; never- 
theless it has been asserted also that the Persian word is derived from 

1 Sui Su, Ch. 83, p. 7 b. 
■ Ch. 185, p. 19. 

* It is not a Tibetan name, as supposed by Roediger and Pott (Z.f. K. Mori., 
Vol. IV, p. 268). 

* These various attempts at spelling show plainly that the term has the status 
of a loan-word, and that the Sanskrit term has nothing to do with the name of the 
people who may have supplied the product, the T&yyavot. in the Himalaya of 
Ptolemy (Yule, Hobson-Jobson, p. 923). How should borax be found in the 
Himalaya ! 

5 Klaproth, M6moires relatifs a l'Asie, Vol. Ill, p. 347. 

6 See, further, T'oung Pao, 1914, pp. 88-89. 

7 D. Hanbury, Science Papers, pp. 217, 276. 


504 Sino-Iranica 

the Chinese. F. de Mely 1 argues that nao-Sa is written ideographically, 
and that the text of the Pen ts'ao kan mu adds, "II vient de la province 
de Chen-si; on le tire d'une montagne d'ou il sort continuellement des 
vapeurs rouges et dangereuses et tres difficile a aborder par rapport a 
ces m£mes vapeurs. II en vient aussi de la Tartarie, on le tire des 
plaines ou il y a beaucoup de troupeaux, de la m£me facon que le 
salpetre de houssage; les Tartares et gens d'au dela de la Chine salent 
les viandes avec ce sel." Hence F. de Mely infers that the Persians, on 
their part, borrowed from the Chinese their nao-Sa, to which they added 
the ending dzer, as in the case of the bezoar styled in Persian badzeher. 2 
The case, however, is entirely different. The term nao-la is written 
phonetically, not ideographically, as shown by the ancient transcription 
iH ^ in the Sui Annals (see below) and the variant Ml ffi (properly 
nun-fa, but indicated with the pronunciation nao-Sa) f also the syno- 
nymes ti yen %K IS ("salt of the barbarians") and Pei-t'in ia At JS ffi 
("ore of Pei-t'ih," in Turkistan), which appear as early as the Sung 
period in the T'u kin pen ts'ao of Su Sun, allude to the foreign origin of 
the product. The term is thus plainly characterized as a foreign loan 
in the Pen ts'ao kan mu. This, further, is brought out by the history of 
the subject. The word is not found in any ancient Chinese records. 
The Chinese learned about nao-la in Sogdiana and Kuca for the first 
time during the sixth century a.d. The Pen ts'ao of the T'ang period is 
the earliest pharmacopoeia that mentions it. Su Kuh H #, the reviser 
of this work, and the author of the Cen lei pen ts'ao, know of but one 
place of provenience, the country of the Western Zuh M ■?& (F. de 
Mely's "Tartary"). It is only Su Sun S£ ^ of the Sung period, who 
in his T'u kin pen ts'ao remarks, "At present it occurs also in Si-lian 
and in the country Hia [Kan-su] as well as in Ho-tuh [San-si], Sen-si, 
and in the districts of the adjoining regions" 'T'M'ilHS^i^^ 
I^H^jlm^Pi^W^ [note the additions of 4* "at present" and 
iff "also"]. And he hastens to add, "However (#S), the pieces coming 
from the Western Zuh are clear and bright, the largest having the size 
of a fist and being from three to five ounces in weight, the smallest 

1 L'Alchimie chez les Chinois (Journal asiatique, 1895, II, p. 338) and Lapidaire 
chinois, p. li. 

2 All this is rather lack of criticism or poor philology. The Persian word in 
question is pazahr, literally meaning "antidote" (see below, p. 525). Neither this 
word nor nusadir has an ending like dzer, and there is no analogy between the two. 

3 According to the Pie pen lu #!] 7^ £fe, cited in the Cen lei pen ts'ao (Ch. 5, 
p. 10, ed. of 1587), the transcription nun-la should represent the pronunciation of 
the Hu people; that is, Iranians. Apparently it was an Iranian dialectic variation 
with a nasalized vowel u. It is indicated as a synonyme of nao-Sa in the Si yao er 
ya of the T'ang period (see Beginnings of Porcelain, p. 115). 

Iranian Minerals — Sal Ammoniac 505 

reaching the size of a finger and being used for medical purposes." 1 
It is accordingly the old experience that the Chinese, as soon as they 
became acquainted with a foreign product, searched for it on their own 
soil, and either discovered it there, or found a convenient substitute. 
In this case, Su Sun plainly indicates that 'the domestic substitute was 
of inferior quality; and there can be no doubt that this was not sal 
ammoniac, which is in fact not found in China, but, as has been demon- 
strated by D. Hanbury, 2 chloride of sodium. As early as the eighteenth 
century it was stated by M. Collas 3 that no product labelled nao-to 
in Peking had any resemblance to our sal ammoniac. 

H. E. Stapleton, 4 author of a very interesting study on the employ- 
ment of sal ammoniac in ancient chemistry, has hazarded an etymo- 
logical speculation as to the term nao-Sa. Persian nuladur appears to 
him to be the Chinese word nau-Sa, suffixed by the Persian word daru 
("medicine"), 5 and the Sanskrit navasdra would also seem to be simply 
the Chinese name in a slightly altered form. H. E. Stapleton is a 
chemist, not a philologist; it therefore suffices to say that these specu- 
lations, as well as his opinion "that the syllables nau-$a appear to be 
capable of complete analysis into Chinese roots," 6 are impossible. 

The Hindustani name can by no means come into question as the 
prototype of the Chinese term, as proposed by F. P. Smith 7 and T. 
Waiters; 8 for the Chinese transcription was framed as early as the 
sixth century a.d., when Hindustani was not yet in existence. The 
Hindustani is simply a Persian loan-word of recent date, as is 
likewise Neo-Sanskrit naiqadala; while Sanskrit navasdra, navasddara, 
or narasdra, the vacillating spelling of which betrays the character 
of a loan-word, is traceable to a more ancient Iranian form (see 

In the Sui iw 9 we meet the term in the form HI ^ nao-Sa, stated to 

1 See also Pen ts'ao yen i, Ch. 6, p. 4 b (ed. of Lu Sin-yuan). 

2 Science Papers, pp. 217, 276. 

3 M6moires concernant les Chinois, Vol. XI, 1786, p. 330. 

* Sal Ammoniac: a Study in Primitive Chemistry (Memoirs As. Soc. Bengal, 
Vol. I, 1905, pp. 40-41). 

5 He starts from the popular etymology nils' daru ("life-giving medicine"), 
which, of course, is not to be taken seriously. 

6 Even if this were the case, it would not tend to prove that the word is of 
Chinese origin. As is now known to every one, there is nothing easier to the Chinese 
than to transcribe a foreign word and to choose such characters as will convey a 
certain meaning. 

7 Contributions toward the Materia Medica of China, p. 190. 

8 Essays on the Chinese Language, p. 350. 
3 Ch. 83, pp. 4 b and 5 b. 

506 Sino-Iranica 

be a product of K'afi (Sogdiana) and Kuca. 1 The fact that this tran- 
scription is identical with fifi we recognize from the parallel passage in 
the Pet U, 2 where it is thus written. The text of the Sui Annals with 
reference to Iranian regions offers several such unusual modes of 
writing, where the Pet H has the simple types subsequently adopted as 
the standard. The variation of the Sui Annals, at all events, demon- 
strates that the question is of reproducing a foreign word; and, since 
it hails from Sogdiana, there can be no doubt that it was a word of the 
Sogdian language of the type *navsa or *naf sa (cf. Sanskrit navasdra, 
Armenian navt\ Greek va<t>da)\ Persian na$ddir, nuladir, nau$ddir, 
nauSddur, noladur, being a later development. It resulted also 
in Russian nulatyr. In my opinion, the Sogdian word is related 
to Persia neft ("naphta"), which may belong to Avestan naptaj 
("moist"). 3 

Tribute-gifts of nao-la are not infrequently mentioned in the Chinese 
Annals. In a.d. 932, Wan Zen-mei 3E £l H, Khan of the Uigur, pre- 
sented to the Court among other objects ta-p'en ia ("borax") 4 and sal 
ammoniac (kan &z). 5 In a.d. 938 Li Sen-wen $ ^ 3&C, king of Khotan, 
offered nao-la and ta-p'en ia ("borax") to the Court; and in a.d. 959 
jade and nao-$a were sent by the Uigur. 6 The latter event is recorded 
also in the Kiu Wu Tat U, 7 where the word is written 01 ffi, pho- 
netically kan-la, but apparently intended only as a graphic variant 
for nao-$a. s The same work ascribes sal ammoniac (written in the same 
manner) to the T'u-fan (Tibetans) and the Tah-hian (a Tibetan tribe 
in the KukunOr region). 9 In the T'ang period the substance was well 

1 According to Masudi (Barbier de Meynard, Les Prairies d'or, Vol. I, p. 347), 
sal-ammoniac mines were situated in Soghd, and were passed by the Moham- 
medan merchants travelling from Khorasan into China. KuSa still yields sal am- 
moniac (A. N. Kuropatkin, Kashgaria, pp. 27, 35, 76). This fact is also noted in 
the Hui k'ian ci (Ch. 2), written about 1772 by two Manchu officials, Fusamb6 
and Surde, who locate the mine 45 It west of Kuca in the Sartatsi Mountains, and 
mention a red and white variety of sal ammoniac. Cf. also M. Reinaud, Relation 
des voyages faits par les Arabes et les Persans dans l'lnde et a la Chine, Vol. I, 


2 Ch. 97, p. 12. 

3 Cf. P. Horn, Neupersische Etymologie, No. 1035; H. H^bschmann, Persische 
Studien, p. 101, and Armen. Gram., p. 100. 

4 As I have shown on a former occasion (T'oung Pao, 1914, p. 88), Chinese 
P'en (*buh) is a transcription of Tibetan bid. 

5 Ts'e fu yuan kwei, Ch. 972, p. 19. 

8 Wu Tax hui yao, Chs. 28, p. 10 b; and Ch. 29, p. 13 b (ed. of Wu yin tien). 
' Ch. 138, p. 3. 

8 The character kan is not listed in K'ah-hi's Dictionary. 

9 Ch. 138, pp. 1 b, 3 a. 

Iranian Minerals — Sal Ammoniac 507 

known. The Si yao er ya 1 gives a number of synonymes of Chinese 
origin, as kin tsei ^ %£,, c'i Sa # ffi ("red gravel"), pat hat tsin 6 M 
Hf ("essence of the white sea"). 

Sal ammoniac is found in Dimindan in the province of Kirman. 
Yaqut (1179-1229) gives after Ibn al-Faqih (tenth century) a descrip- 
tion of how nuladir is obtained there, which in the translation of C. 
Barbier de Meynard 2 runs as follows: — 

"Cette substance se trouve principalement dans une montagne 
nomm6e Donbawend, dont la hauteur est eValue"e a 3 farsakhs. Cette 
montagne est a 7 farsakhs de la ville de Guwa&r. On y voit une caverne 
profonde d'ou s'^chappent des mugissements semblables a ceux des 
vagues et une fumee £paisse. Lorsque cette vapeur, qui est le principe 
du sel ammoniac, s'est attachee aux parois de l'orifice, et qu'une certaine 
quantity s'est solidifiee, les habitants de la ville et des environs viennent 
la recueillir, une fois par mois ou tous les deux mois. Le sulthan y envoie 
des agents qui, la recolte faite, en prel event le cinquieme pour le tr6sor; 
les habitants se partagent le reste par la voie du sort. Ce sel est celui 
qu'on expe"die dans tous les pays." 

Ibn Haukal describes the mines of SetruSteh thus: 8 "The mines 
of sal ammoniac are in the mountains, where there is a certain cavern, 
from which a vapor issues, appearing by day like smoke, and by night 
like fire. Over the spot whence the vapor issues, they have erected a 
house, the doors and windows of which are kept so closely shut and 
plastered over with clay that none of the vapor can escape. On the 
upper part of this house the copperas rests. When the doors are to be 
opened, a swiftly-running man is chosen, who, having his body covered 
over with clay, opens the door; takes as much as he can of the copperas, 
and runs off; if he should delay, he would be burnt. This vapor comes 
forth in different places, from time to time; when it ceases to issue from 
one place, they dig in another until it appears, and then they erect that 
kind of house over it; if they did not erect this house, the vapor would 
burn, or evaporate away." 

Taxes are still paid in this district with sal ammoniac. Abu Mansur 
sets forth its medicinal properties. 4 

1 See Beginnings of Porcelain (this volume, p. 1 15). 

2 Dictionnaire g£ographique de la Perse, p. 235 (Paris, 1861). Ibn al-Faqlh's 
text is translated by P. Schwarz (Iran im Mittelalter, p. 252). According to Ibn 
Haukal (W. Ouseley, Oriental Geography of Ebn Haukal, p. 233), sal-ammoniac 
mines were located in Maweralnahr (Transoxania). 

3 W. Ouseley, op. tit., p. 264. 

4 Achundow, Abu Mansur, p. 144. — Abel-Remusat (Melanges asiatiques, 
Vol. I, p. 209, 1825), translating from the Japanese edition of the cyclopaedia San 
ts'ai t'u hut, gave the following interesting account: "Le sel nomme' (en chinois) 

508 Sino-Iranica 

The Tibetans appear to have received sal ammoniac from India, as 
shown at least by their term rgya ts'wa ("Indian salt"), literally trans- 
lated into Mongol Andtkdk dabusu. Mongol Andtkdk is a reproduction 
of Chinese *In-duk-kwok ("country of India"). The informants of 
M. Collas 1 stated that the nao-ia of the Peking shops came from Tibet 
or adjacent places. Lockhart received in Peking the information that 
it is brought from certain volcanic springs in Se-c'wan and in Tibet. 1 

80. $? $£ $£ mi-Vo-seh, *m / it(m / ir)-da-sah, and $t ^ fl^ mu-to- 
seh t *mut(mur)-ta-sah, litharge, dross of lead, is an exact reproduction 
of Persian mirddsang or murddsang of the same meaning. 3 Both tran- 
scriptions are found in the Pen ts'ao of the T'ang dynasty, written 
about the middle of the seventh century. 4 Therefore we are entitled to 
extend the Persian word into the period of Middle Persian. Su Kun, 
the reviser of the T'an pen ts'ao, states expressly that both mi-Vo and 
mu-to are words from the language of the Hu or Iranians (#J B -&), 
and that the substance comes from or is produced in Persia, being in 
shape like the teeth of the yellow dragon, but stronger and heavier; 
there is also some of white color with veins as in Yun-nan marble. Su 
Sun of the Sung period says that then ("at present") it was also found 

nao-cha (en persan nouchader) et aussi sel de Tartarie, sel volatil, se tire de deux 
montagnes volcaniques de la Tartarie centrale; l'une est le volcan de Tourfan, qui 
a donne' a cette ville (ou pour mieux dire a une ville qui est situ£e a trois lieues de 
Tourfan, du c6t6 de Test) le nom de Ho-tcheou, ville de feu; l'autre est la montagne 
Blanche, dans le pays de Bisch-balikh; ces deux montagnes jettent continuellement 
des flammes et de la fum6e. II y a des cavit6s dans lesquelles se ramasse un liquide 
verdatre. Expose" a l'air, ce liquide se change en un sel, qui est le nao-cha. Les 
gens du pays le recueillent pour s'en servir dans la preparation des cuirs. Quant a 
la montagne de Tourfan, on en voit continuellement sortir une colonne de fum£e; 
cette fumee est remplac£e le soir par une flamme semblable a celle d'un flambeau. 
Les oiseaux et les autres animaux, qui en sont 6clair6s, paraissent de couleur rouge. 
On appelle cette montagne le Mont-de-Feu. Pour aller chercher le nao-cha, on met 
des sabots, car des semelles de cuir seraient trop vite bruises. Les gens du pays 
recueillent aussi les eaux-meres qu'ils font bouillir dans des chaudieres, et ils en 
retirent le sel ammoniac, sous la forme de pains semblables a ceux du sel commun. 
Le nao-cha le plus blanc est r6put6 le meilleur; la nature de ce sel est tres-p£n6trante. 
On le tient suspendu dans une po&le au-dessus du feu pour le rendre bien sec; on y 
ajoute du gingembre pour le conserver. Expose au froid ou a l'humidit6, il tombe en 
deliquescence, et se perd." Wan Yen-te, who in a.d. 981 was sent by the Chinese 
emperor to the ruler of Kao-S'ah, was the first to give an account of the sal-ammoniac 
mountain of Turkistan (Bretschneider, Mediaeval Researches, Vol. II, p. 190). 
See also P. de Mely, Lapidaire chinois, p. 140; W. Schott, Zur Uigurenfrage, II, 
p. 45 (Abh. Berl. Akad., 1875) and Ueber ein chinesisches Mengwerk (ibid., 1880, 
p. 6); Geerts, Produits, p. 322. 

1 Memoires concernant les Chinois, Vol. XI, p. 331. 

2 D. Hanbury, Science Papers, p. 277. 

1 Cf. Hubschmann, Armen. Gram., p. 270. 

4 Cen lei pen ts'ao, Ch. 4, p. 31; and Pen ts'ao kan mu, Ch. 8, p. 8 b. 

Iranian Minerals — Litharge, Gold 509 

in the silver and copper foundries of Kwan-tun and Fu-kien. It is 
further mentioned briefly in the Pen ts'ao yen ioim 6, 1 which maintains 
that the kind with a color like gold is the best. 

According to Yaqut, mines of antimony, known under the name 
razi, litharge, lead, and vitriol, were in the environs of Donbawend or 
Demawend in the province of Kirman. 2 In the Persian pharmacopoeia 
of Abu Mansur, the medicinal properties of litharge are described under 
the Arabicized name murddsanj, to which he adds the synonymous term 
murtak* Pegoletti, in the fourteenth century, gives the word with a 
popular etymology as morda sangue* The Dictionary of Four Lan- 
guages 5 correlates Chinese mi-Vo-sen with Tibetan gser-zil (literally, 
"gold brightness"), 6 Manchu lirlan, and Mongol jildunur. 7 

81. Palladius 8 offers a term $£ M & tse-mo kin with the meaning 
"gold from Persia," no source for it being cited. In the Pen ts'ao kan 
mu* the tse-mo kin of Po-se (Persia) is given as the first in a series of 
five kinds of gold of foreign countries, 10 without further explanation. 
The term occurs also in Buddhist literature: Chavannes 11 has found it 
in the text of a Jataka, where he proposes as hypothetical translation, 
"un amas d'or raffine" rouge." It therefore seems to be unknown what 
the term signifies, although a special kind of gold or an alloy of gold is 
apparently intended. The Swi kin ?m ^S & 12 says that the first 
quality of gold, according to Chinese custom, is styled tse-mo kin 
(written as above); according to the custom of the barbarians, how- 
ever, yan-mai 81 }§. From this it would appear that tse-mo is a Chinese 
term, not a foreign one. 

1 Ch. 5, p. 6 b (ed. of Lu Sin-yuan). 

2 Barbier de Meynard, op. tit., p. 237. 

I Achundow, Abu Mansur, p. 139. This form goes back to Middle Persian 
murtak or martak. 

* Yule, Cathay, new ed., Vol. Ill, p. 167. 

s Ch. 22, p. 71. 

8 Jaeschke, in his Tibetan Dictionary, was unable to explain this term. 

7 Kovalevski, in his Mongol Dictionary, explains this word wrongly by 

8 Chinese-Russian Dictionary, Vol. II, p. 203. 

9 Ch. 8, p. 1 b. 

10 The four others are, the dark gold of the eastern regions, the red gold of 
Lin-yi, the gold of the Si-zuh, and the gold of Can-5'en (Camboja). The five kinds 
of foreign gold are mentioned as early as the tenth century in the Pao ts'an lun 


II Fables et contes de lTnde, in Actes du XIV 8 Congres des Orientalistes, 
Vol. I, 1905, p. 103. 

12 Ch. 36, p. 18 b (ed. Wu-c'an, 1877). See p. 622. 

510 Sino-Iranica 

The Ko ku yao lun 1 has a notice of tse kin f& 4£ ("purple gold") 
as follows: "The ancients say that the pan-lian *F M money 2 is tse 
kin. The people of the present time make it by mixing copper with 
gold, but our contemporaries have not yet seen genuine tse kin." 
The same alloy is mentioned as a product of Ma-k'o-se-li in the 
Tao i Si lio, written in 1349 by Wan Ta-yuan. 3 I am not sure, of 
course, that this tse kin is identical with tse-mo kin. 

In the same manner as the Chinese speak of foreign gold, they also 
offer a series of foreign silver. There are four kinds; namely, silver of 
Sin-ra (in Korea), silver of Po-se (Persia), silver of Lin-yi, and silver 
of Yiin-nan. Both gold and silver are enumerated among the products 
of Sasanian Persia. The Hai yao pen ts'ao cites the Nan yiie li of the 
fifth century to the effect that the country Po-se possesses a natural 
silver-dust $& M , employed as a remedy, and that remedies are tested 
by means of finger-rings.* Whether Persia is to be understood here 
seems doubtful to me. Gold-dust is especially credited to the country 
of the Arabs. 5 

82. 1ft $& yen-lii ("the green of salt," various compositions with 
copper-oxide) is mentioned as a product of Sasanian Persia 6 and of 
Kuca. 7 Su Kuh of the T'ang (seventh century) points it out as a product 
of Karasar (Yen-£i M w), found in the water on the lower surface of 
stones. Li Sun, who wrote in the second half of the eighth century, 
states that "it is produced in the country Po-se (Persia) adhering to 
stones, and that the kind imported on ships is called H-lii ^ $&('the 
green of the stone ') ; its color is resistant for a long time without chang- 
ing; the imitation made in China from copper and vinegar must not 
be employed in the pharmacopoeia, nor does it retain its color long." 
Li Si-cen employs the term "green salt of Po-se." 8 The substance was 
employed as a remedy in eye-diseases. 

This is Persian zingar (Arabic zinjar), described in the stone-book 
of Pseudo-Aristotle as a stone extracted from copper or brass by means 

1 Ch. 6, p. 12 b. 

2 See Beginnings of Porcelain, p. 83. 

8 Rockhill, T'oung Pao, 1915, p. 622. 
* Cen lei pen ts'ao, Ch. 4, p. 23. 

5 Ibid., Ch. 4, p. 21 b. 

6 Sui Su, Ch. 83, p. 7 b. 

7 Cou Su, Ch. 50, p. 5; Sui Su, Ch. 83, p. 5 b. 

8 Cf. also Geerts, Produits, p. 634; F. de Mely, Lapidaire chinois, pp. 134, 
243. According to Geerts, the term is applied in Japan to acetate of copper, formerly 
imported, but now prepared in the country. 

Iranian Minerals — Copper-Oxides, Salt, Zinc 511 

of vinegar, and employed as an ingredient in many remedies for eye- 
diseases. 1 

83. The Emperor Yan (a.d. 605-616) of the Sui dynasty, after 
his succession to the throne, despatched Tu Han-man tt fr ffil to the 
Western Countries. He reached the kingdom of Nan 52c (Bukhara), 
obtained manicolored salt (wu se yen), and returned. 2 Istaxri relates 
that in the district of Darabejird there are mountains of white, yellow, 
green, black, and red salts; the salt in other regions originates from the 
interior of the earth or from water which forms crystals; this, however, 
is salt from mountains which are above the ground. Ibn Haukal adds 
that this salt occurs in all possible colors. 3 

The Pei hu lu* distinguishes red, purple, black, blue, and yellow 
salts. C'i yen 7$ |ft ("red salt ") like vermilion, and white salt like jade, 
are attributed to Kao-S'an (Turfan). 5 Black salt (hei yen) was a product 
of the country Ts'ao (Jaguda) north of the Ts'un-lin. 6 It is likewise 
attributed to southern India. 7 These colored salts may have been im- 
pure salt or minerals of a different origin. 

84. ffe 35 Vou-H is mentioned as a metallic product of Sasanian 
Persia (enumerated with gold, silver, copper, pin, iron, and tin) in the 
Sui iw. 8 It is further cited as a product of Nu kwo, the Women's Realm 
south of the Ts'un-lih; 9 of A-lo-yi-lo K & & $k in the north of Utfdi- 
yana, 10 and of the Arabs (Ta-si). 11 Huan Tsan's Memoirs contain the 
term three times, once as a product found in the soil of northern India 
(together with gold, silver, copper, and iron) , and twice as a material 
from which Buddhist statues were made. 12 According to the Kin Vu 

1 J. Ruska, Steinbuch des Aristoteles, p. 182; and Steinbuch des Qazwlnl, 
P- 25. 

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

I P. Schwarz, Iran, p. 95. 

4 Ch. 2, p. 11 (ed. of Lu Sin-yuan). 

5 Sui Su, Ch. 83, p. 3 b. In the T'ai p'in hwan yil ki (Ch. 180, p. n b) the same 
products are assigned to Ku-§i 3f£ ^fp (Turfan). 

8 Sui $u, Ch. 83, p. 8. 

7 T'an Su, Ch. 221 A, p. 10 b. 
* Ch. 83, p. 7 b. 

9 T'ai p'in hwan yil ki, Ch. 186, p. 9. 

10 Ibid., p. 12 b. 

II Ibid., p. 15 b. 

12 Cf. S. Julien, Memoires sur les contrees occidentales, Vol. I, pp. 37, 189, 
354. Julien is quite right in translating the term by laiton ("brass"). Palladius 
(Chinese-Russian Dictionary, Vol. II, p. 16) explains it as "brass with admixture of 
lead, possessing attractive power." The definition of Giles ("rich ore brought 
from Persia supposed to be an ore of gold and copper, or bronze") is inexact. T'ou- 

512 Sino-Iranica 

swi H hi M st£ Wl B^ Hfi, written in the sixth century, the needles used 
by women on the festival of the seventh day of the seventh month 1 
were made of gold, silver, or Vou-U. 2 Under the T'ang, Vou-H was an 
officially adopted alloy, being employed, for instance, for the girdles of 
the officials of the eighth and ninth grades. 3 It was sent as tribute 
from Iranian regions; for instance, in a.d. 718, from Maimargh (north- 
west of Samarkand). 4 

The Ko ku yao lun states, " T % ou-U is the essence of natural copper. 
At present zinc-bloom is smelted to make counterfeit Vou. According to 
Ts'ui Fan -Hi fr§, one catty of copper and one catty of zinc-bloom will 
yield Vou-U. The genuine Vou is produced in Persia. It looks like gold, 
and, when fired, assumes a red color which will never turn black." 
This is clearly a description of brass which is mainly composed of copper 
and zinc. Li Si-Sen 5 identifies Vou-U with the modern term hwah Vuh 
("yellow copper"); that is, brass. According to T'an Ts'ui, 6 Vou-H is 
found in the C'6-li $ M t'u-se of Yun-nan. 

The Chinese accounts of Vou or Vou-H agree with what the Persians 
and Arabs report about tutiya. It was in Persia that zinc was first mined, 
and utilized for a new copper alloy, brass. Ibn al-Faqih, who wrote 
about a.d. 902, has left a description of the zinc-mines situated in a 
mountain Dunbawand in the province of Kirman. The ore was (and 
still is) a government monopoly. 7 Jawbari, who wrote about 1225, has 
described the process of smelting. 8 The earliest mention of the term 
occurs in the Arabic stone-book of Pseudo-Aristotle (ninth century), 9 
where the stone tutiya is explained as belonging to the stones found in 
mines, with numerous varieties which are white, yellow, and green; 

Si is only said to resemble gold, and the notion that brass resembles gold turns up in 
all Oriental writers. See also Beal, Records of the Western World, Vol. I, p. 51; 
and Chavannes (T'oung Pao, 1904, p. 34), who likewise accepts the only admissible 
interpretation, ' ' brass. ' ' 

1 Cf. W. Grube, Zur Pekinger Volkskunde, p. 76; J. Przyluski, T'oung Pao, 
1914, p. 215. 

2 P'ei wen yiln fu, Ch. 100 A, p. 25. 

' Jade, p. 286; cf. also Ta T'an leu Hen, Ch. 8, p. 22. 

4 Chavannes, T'oung Pao, 1904, p. 34. 

6 Pen ts'ao kan mu, Ch. 8, pp. 3 and 4. Cf. also Geerts, Produits, p. 575. 

6 Tien hai yu hen li, Ch. 2, p. 3 b. 

7 P. Schwarz, Iran im Mittelalter, p. 252. 

8 G. Ferrand, Textes relatifs a l'ExtrSme-Orient, p. 610 (cf. also pp. 225, 228; 
and Leclerc, Traits des simples, Vol. I, p. 322). 

9 J. Ruska, Steinbuch des Aristoteles, p. 175. J. Beckmann (Beytrage zur 
Geschichte der Erfindungen, Vol. Ill, p. 388) states that the word first occurs in 
Avicenna of the eleventh century. 

Iranian Minerals — Zinc 513 

the quarries are located on the shores of Hind and Sind. This is prob- 
ably intended for vitriol or sulphate of copper. 1 

In Chinese fou-U, the second element H ("stone") does not form 
part of the transcription; the term means simply "Vou stone," and Vou 
(*tu) reproduces the first syllable of Persian tutiya, which, on the basis 
of the Sui Annals, we are obliged to assign also to the Middle-Persian 
language. To derive the Chinese word from Turkish tiij, as proposed 
by Watters, 2 and accepted without criticism by Hirth, 3 is utterly im- 
possible. The alleged Turkish word occurs only in Osmanli and other 
modern dialects, where it is plainly a Persian loan-word, but not in 
Uigur, as wrongly asserted by Hirth. This theory seems to imply that 
the element Si should form part of the transcription; this certainly is 
out of the question, as 3J represents ancient *§ek or *sak, *zak, and 
could not reproduce a palatal. For the rest, the Chinese records point 
to Iran, not to the Turks, who had no concern whatever with the 
whole business. 4 Two variations of the Persian word have penetrated 
into the languages of Europe. The Arabs carried their tutiya into 
Spain, where it appears as atutia with the Arabic article; in Portuguese 
we have tutia, in French tutie, in Italian tuzia, in English tutty. A final 
palatal occurs in the series Osmanli tuj or tuni, Neo-Greek tovvt£i, 
Albanian tut, Serbian and Bulgarian tut, Rumanian tuciu. Whether 
Sanskrit tuttha, as has been assumed, is to be connected with the Per- 
sian word, remains doubtful to me: the Sanskrit word refers only to 
green or blue vitriol. 5 It is noteworthy that Persian birinj ("brass"), a 
more recent variant of pirin (Kurd pirinjok, Armenian plinj), 6 has not 
migrated into any foreign language, for I am far from being convinced 
that our word "bronze" should be traceable to this type. 7 

The Japanese pronunciation of 1^ 5 is tuseki. The Japanese used 

1 A curious error occurs in Feldhaus' Technik (col. 1367), where it is asserted, 
"Qazwlnl says about 600 that zinc is known in China, and could also be made 
flexible there." Qazwlnl wrote his cyclopaedia in 1 134, and says nothing about 
zinc in China (cf. Ruska, Steinbuch des Qazwlnl, p. 11); but he mentions a tutiya 
mine in Spain (G. Jacob, Studien in arabischen Geographen, p. 13). 

s Essays on the Chinese Language, p. 359. 

1 Chau Ju-kua, p. 81. Tou-Si does not mean "white copper" in the passage 
under notice, but means "brass." "White copper" is a Chinese and quite different 
alloy (see below, p. 555). 

* It is likewise odd to connect Italian tausia (properly taunia) and German 
tauschieren with this word. This is just as well as to derive German tusche from 
an alleged Chinese fuse (Hirth, Chines. Studien, p. 226). 

6 P. C. Ray, History of Hindu Chemistry, 2d ed., Vol. II, p. 25. 

6 Hubschmann, Persische Studien, p. 27. 

7 O. Schrader, Sprachvergleichung und Urgeschichte, Vol. II, p. 73. 

514 Sino-Iranica 

to import the alloy from China, and their HonzO (Pen ts*ao) give for- 
mulas for its preparation. 1 The Koreans read the same word not or 
not-st. The French missionaries explain it as "composition de differents 
m£taux qui sert a faire les cuilleres, etc. Airain, cuivre jaune (premiere 
quality). Cuivre rouge et plomb." 2 

The history of zinc in the East is still somewhat obscure; at least, 
it so appears from what the historians of the metal have written about 
the subject. I quote from W. R. Ingalls: 3 "It is unknown to whom is 
due the honor of the isolation of zinc as a metal, but it is probable that 
the discovery was first made in the East. In the sixteenth century zinc 
was brought to Europe from China and the East Indies under the name 
of tutanego (whence the English term tutenegue), and it is likely that 
knowledge of it was obtained from that source at an earlier date. . . . 
The production of zinc on an industrial scale was first begun in England; 
it is said that the method applied was Chinese, having been introduced 
by Dr. Isaac Lawson, who went to China expressly to study it. In 1740 
John Champion erected works at Bristol and actually began the manu- 
facture of spelter, but the production was small, and the greater part 
used continued to come from India and China." The fact that in the 
eighteenth century the bulk of zinc which came to Europe was shipped 
from India is also emphasized by J. Beckmann, 4 who, writing in 1792, 
regretted that it was then unknown where, how, and when this metal 
was obtained in India, and in what year it had first been brought over 
to Europe. According to the few notices of the subject, he continues, it 
originates from China, from Bengal, from Malakka, and from Malabar, 
whence also copper and brass are obtained. On the other hand, W. 
Ainslie 5 states that by far the greater part of zinc which is met with 
in India is brought from Cochin-China or China, where both the cala- 
mine and blende are common. Again, S. Julien 6 informs us that zinc 
is not mentioned in ancient books, and appears to have been known in 
China only from the beginning of the seventeenth century. 

W. Hommel 7 pleaded for the origin of zinc-production in India, 
whence it was obtained by the Chinese. He does not know, of course, 
that there is no evidence for such a theory in Chinese sources. The 

1 Geerts, Produits, p. 641; F. de Mely, Lapidaire chinois, p. 42. 

2 Dictionnaire cor£en-francais, p. 291. 

1 Production and Properties of Zinc, pp. 2-3 (New York and London, 1902). 

4 Op. cit., Vol. Ill, p. 408. 

s Materia Indica, Vol. I, p. 573. 

8 Industries de l'empire chinois, p. 46. 

7 Chemiker-Zeitung, 1912, p. 905. 

Iranian Minerals — Zinc, Steel 515 

Indian hypothesis, I believe, has been accepted by others. In my opin- 
ion, the art of zinc-smelting originated neither in India nor in China, but 
in Persia. We noted from Ibn al-Faqlh that the zinc-mines of Kirman 
were wrought in the tenth century; and the early Chinese references to 
Vou-li would warrant the conclusion that this industry was prominent 
under the Sasanians, and goes back at least to the sixth century. 

Li Si-cen 1 states that the green copper of Persia can be wrought into 
mirrors. I have no other information on this metal. 

85. 4& or $& £e§ pin Vie, pin iron, is mentioned as a product of Sa- 
sanian Persia, 2 also ascribed to Ki-pin (Kashmir). 3 Mediaeval authors 
like C'ah Te mention it also for India and Hami. 4 The Ko ku yao lun 5 
says that pin Vie is produced by the Western Barbarians (Si Fan), and 
that its surface exhibits patterns like the winding lines of a conch or 
like sesame-seeds and snow. Swords and other implements made from 
this metal are polished by means of gold threads, and then these pat- 
terns become visible; the price of this metal exceeds that of silver. This 
clearly refers to a steel like that of Damascus, on which fine dark lines 
are produced by means of etching acids.' 

Li Si-cen 7 states that pin Vie is produced by the Western Barbarians 
(Si Fan), and cites the Pao is' an lun H M tfc, by Hien Yuan-su 
$f It 2fc of the tenth century, to the effect that there are five kinds of 
iron, one of these being pin Vie, which is so hard and sharp that it can 
cut metal and hard stone. K'ah-hi's Dictionary states that pin is 
wrought into sharp swords. Previous investigators have overlooked the 
fact that this metal is first mentioned for Sasanian Persia, and have 
merely pointed to the late mediaeval mention in the Sung Annals. 8 

The word pin has not yet been explained. Even the Pan-Turks have 
not yet discovered it in Turkish. It is connected with Iranian *spaina, 
Pamir languages spin, Afghan dspina or dspana, Ossetic qfsan. 9 The 

1 Pen ts'ao kan mu, Ch. 8, p. 3 b. 
1 Cou Su, Ch. 50, p. 6; Sui Su, Ch. 83, p. 7 b. 
1 T'ai p'in hwan yu ki, Ch. 182, p. 12 b. 

4 Bretschneider, Mediaeval Researches, Vol. I, p. 146; Kwan yU ki, Ch. 24, 
p. 5 b. 

6 Ch. 6, p. 14 b (ed. of Si yin huan ts'un Su). 

6 A reference to pin Vie occurs also in the San kit sin hwa |Xj ^§ ^f IS. written 
by Yah Yu ^ Iff} in 1360 (p. 19, ed. of Ci pu tsu lai ts'un Su). 

7 Pen ts'ao kan mu, Ch. 8, p. 1 1 b. 

8 Bretschneider, On the Knowledge possessed by the Chinese of the Arabs, 
p. 12, and China Review, Vol. V, p. 21; W. F. Mayers, China Review, Vol. IV, 
P- 175. 

• Hubschmann, Persische Studien, p. 10. 

516 Sino-Iranica 

character pin has been formed ad hoc, and, as already remarked by- 
Mayers, is written also without the classifier; that is, in a purely pho- 
netic way. 

86. fi&i& se-se, *sit-sit (Japanese Htsu-Htsu), hypothetical restora- 
tion *sirsir, a precious stone of Sasanian Persia, which I have discussed 
at some length in my "Notes on Turquois in the East" (pp. 25-35, 
45-55, 67-68). For this reason only a brief summary is here given, with 
some additional information and corrections. I no longer believe that 
se-se might be connected with Shignan (p. 47) or Arabic jaza (p. 52), but 
am now convinced that se-se represents the transcription of an Iranian 
(most probably Sogdian) word, the original of which, however, has not 
yet been traced. Chinese records leave us in the dark as to the character 
of the Iranian se-se. It is simply enumerated in a list of precious stones 
of Persia and Sogdiana (K'an). 1 The T'ang Annals locate the se-se mines 
to the south-east of the Yaxartes in Sogdiana; 2 and the stones were 
traded to China by way of Khotan. 3 Possibly the Nestorians were 
active in bringing to China these stones which were utilized for the 
decoration of their churches. The same history ascribes columns of 
se-se to the palaces of Fu-lin (Syria) ; 4 in this case the question is of a 
building-stone. In ancient Tibet, se-se formed part of the official costume, 
being worn by officials of the highest rank in strings suspended from 
the shoulder. The materials ranking next to this stone were gold, 
plated silver, silver, and copper, 5 — a clear index of the fact that se-se 
was regarded in Tibet as a precious stone of great value, and surpassing 
gold. The Tibetan women used to wear beads of this stone in their 
tresses, and a single bead is said to have represented the equivalent of 
a noble horse. 6 Hence arose the term ma kia lu E$ HI % ("pearl or bead 
equalling a horse in price"). These beads are treated in the Ko ku yao 
lun 7 as a separate item, and distinct from turquois. 8 

In the T'ang period, se-se stones were also used as ornaments by the 

1 Pei li, Ch. 97, pp. 7 b, 12; Cou lu, Ch. 50, p. 6; Sui lu, Ch. 83, p. 7 b; Wei lu, 
Ch. 102, pp. 5 a, 9 b. 

2 T'an lu, Ch. 221 B, p. 2 b. 
» T'an lu, Ch. 221 A, p. 10 b. 

4 Kiu T'an lu, Ch. 198, p. 11 b; T'an lu, Ch. 221 b, p. 7 b. 

5 T'an lu, Ch. 216 A, p. 1 b (not in Kiu T'an lu). 
8 Sin Wu Tai li, Ch. 74, p. 4 b. 

7 Ch. 6, p. 5 b. 

8 As justly said by Geerts (Produits de la nature japonaise et chinoise, p. 481), 
it is possible that ma kia lu (Japanese bakalu) is merely a synonyme of the emerald. 
Also in the Pen ts'ao kan mu (Ch. 8, p. 17 b) a distinction is made between the two 
articles, tien-tse ^ J* being characterized as pi$ek, ma kia cu as ts'ui 2$.. 

Iranian Precious Stones — Se-se 517 

women of the Nan Man (the aboriginal tribes of southern China), being 
fastened in their hair; 1 and were known in the kingdom of Nan-cao. 2 
Likewise the women of Wei-cou $1 ')'H in Se-c'wan wore strung se-se 
in their hair. 3 Further, we hear at the same time of se-se utilized by the 
Chinese and even mined in Chinese soil. In some cases it seems that 
a building-stone is involved; in others it appears as a transparent 
precious stone, strung and used for curtains and screens, highly valued, 
and on a par with genuine pearls and precious metals. 4 Under the year 
786, the T'ang Annals state, "The Kwan-c'a-si $tt^$l 5 of San-cou 
^ ffi (in Ho-nan), Li Pi ^ ^ by name, reported to the throne that the 
foundries of Mount Lu-si M. J£ produce se-se , and requested that it 
should be prohibited to accept these stones in the place of taxes; where- 
upon the Emperor (Te Tsuh) replied, that, if there are se-se not pro- 
duced by the soil, they should be turned over to the people, who are 
permitted to gather them for themselves." The question seems to be 
in this text of a by-product of metallic origin; and this agrees with what 
Kao Se-sun remarks in his Wei lio, that the se-se of his time (Sung period) 
were made of molten stone. 

I have given two examples of the employment of se-se in objects of 
art from the K'ao ku Vu and Ku yii t'u p'u (p. 31). Meanwhile I have 
found two instances of the use of the word se-se in the Po ku Vu lu, 
published by Wan Fu in 1107-11. In one passage of this work, 6 the 
patina of a tin JR, attributed to the Cou period, is compared with the 
color of se-se: since patinas occur in green, blue, and many other hues, 
this does not afford conclusive evidence as to the color of se-se. In 
another case 7 a small tin dated in the Han period is described as being 
decorated with inlaid gold and silver, and decorated with the seven 
jewels (saptaratna) and se-se of very brilliant appearance. This is 
striking, as se-se are not known to be on record under the Han, but first 
appear in the accounts of Sasanian Persia: either the bronze vessel in 
question was not of the Han, but of the T'ang; or, if it was of the Han, 
the stone thus diagnosed by the Sung author cannot have been identical 
with what was known by this name under the T'ang. I already had 
occasion to state (p. 33) that the Sung writers knew no longer what the 

1 T'an $u, Ch. 222 A, p. 2. 1 

5 Man $u, p. 48. 

3 T'ai p'in hwan yu ki, Ch. 78, p. 9 b. 

* Min hwan tsa lu, Ch. B, p. 4; Wei lio, Ch. 5, p. 3; Tu yan tsa pien, Ch. A, pp. 3, 
8; Ch. c, pp. 5, 9 b, 14 b. 

5 Official designation of a Tao-t'ai. 

6 Ch. 3, p. 15 b. 

7 Ch. 5, p. 46 b. 

518 Sino-Iranica 

se-se of the T'ang really were, that the T'ang se-se were apparently 
lost in the age of the Sung, and that substitutes merely designated by 
that name were then in vogue. 

Under the Yuan or Mongol dynasty the word se-se was revived. 
C'afi Te, the envoy who visited Bagdad in 1259, reported se-se among 
the precious stones of the Caliph, together with pearls, lapis lazuli, and 
diamonds. A stone of small or no value, found in Kin-cou (in Sen-kin, 
Manchuria), was styled se-se; 1 and under the reign of the Emperor 
C'eh-tsuh (1 295-1307) we hear that two thousand five hundred catties 
of se-se were palmed off on officials in lieu of cash payments, a practice 
which was soon stopped by imperial command. 2 Under the Ming, se-se 
was merely a word vaguely conveying the notion of a precious stone of 
the past, and transferred to artifacts like beads of colored glass or 
clay. 3 

The Chinese notices of se-se form a striking analogy to the accounts 
of the ancients regarding the emerald (smaragdos) , which on the one 
hand is described as a precious stone, chiefly used for rings, on the 
other hand as a building-stone. Theophrastus 4 states, "The emerald 
is good for the eyes, and is worn as a ring-stone to be looked at. It is 
rare, however, and not large. Yet it is said in the histories of the 
Egyptian kings that a Babylonian king once sent as a gift an emerald 
of four cubits in length and three cubits in width; there is in the temple 
of Jupiter an obelisk composed of four emeralds, forty cubits high, four 
cubits wide, and two cubits thick. The false emerald occurs in well- 
known places, particularly in the copper-mines of Cyprus, where it 
fills lodes crossing one another in many ways, but only seldom is it 
large enough for rings." H. O. Lenz 5 is inclined to understand by the 
latter kind malachite. Perhaps the se-se of Iran and Tibet was the 
emerald; the se-se used for pillars in Fu-lin, malachite. No Chinese 
definition of what se-se was has as yet come to light, and we have to 
await further information before venturing exact and positive identifi- 

In Buddhist literature the emerald appears in the transcription 
mo-lo-k'te-Vo IP $1 ffrU P2, 6 corresponding to Sanskrit marakata. In the 
transcription fifr /fc M cu-mu-la, in the seventeenth century written 
JjJH. # $& tsu-mu-lii, the emerald appears to be first mentioned in the 

1 Yuan si, Ch. 24, p. 2 b. 

2 Ibid., Ch. 21, p. 7 b. 

* Cf. Notes on Turquois, p. 34. 

4 De lapidibus, 42. 

6 Mineralogie der Griechen und Romer, p. 20. 

8 Fan yi min yi tsi, Ch. 8, p. 14 b. 

Iranian Precious Stones — Emerald, Turquois 519 

Co ken lu, written in 1366. l The Dictionary in Four Languages 2 writes 
this word Isie-mu-lu M !• Hsk. This is a transcription of Persian 

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 bdrkat, in 
Syriac borko, in Arabic zummurud, in Armenian zemruxt; 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 Nlsapur and Kirman, is first mentioned under the name 
tien-tse ^} -? 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 ;§! 5ifc represents a transcrip- 
tion of Persian firuza ("turquois"), as proposed by Waiters 6 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- 
S'wan, Yun-nan. 8 The Geography of the Ming dynasty indicates a 
turquois-mine in Nan-nin cou :£: ^ ^H 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-lii, 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 lii pao Si jSjsjfc |!£ ^j ("green precious stone"); 
see Geerts, Produits, p. 481. 

1 Ch. 22, p. 66. 

' 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. 

8 As intimated by me in American Anthropologist, 1916, p. 589. Tien-tse as the 
product of Pan-ta-li are mentioned in the Too i li 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 lii ^ 
§| jj^, 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 Sino-Iranica 

Yun-nan Province. 1 In this text, the term pi t'ien-tse §1 *% -?" is em- 
ployed. T'an Ts'ui 2 says that turquoises {pi Vien) are produced in the 
Moh-yah t'u-se jIl ^ i. ^ of Yun-nan. In the Kin-nan fu U % 5: 
$f J^, 3 the gazetteer of the prefecture of Hin-nan in southern Sen-si, 
it is said that pi Vien (written $l|) 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. 4£ >fit kin tsin ("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-ZH SB ^>. 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 H 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 lan-l % i 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 lajvard (Arabic lazvard). Another Arabic word is linej, 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 Vun li, Ch. 86, p. 8. 

2 Tien hai yu hen ci, 1799, Ch. 1, p. 6 b (ed. of Wen yin lou yu ti ts'un Su). See 
above, p. 228. T'u-se are districts under a native chieftain, who himself is subject to 
Chinese authority. 

3 Ch. n, p. 11 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'Ecolefrangaise, Vol. VIII, p. 372), where the question 
is of coral and turquois used by the Ku-tsuh (a Tibetan tribe) women as ornaments; 
instead of yuan-song, as there transcribed, read lii sun Si ^ %k ^. 

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, Traits 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 l'Azur and Lazuli) have been, like the ruby mines, 
celebrated for ages. They lie 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 1838. 2 
The produce now is said to be of very inferior quality, and in quantity 
from thirty to sixty pud (thirty-six lbs. each) annually. The best 
quality sells at Bokhara at thirty to sixty tillas, or 12 /. to 24 /. the pud 
(Manphul)." 3 In the Dictionary of Four Languages, 4 lapis lazuli is 
styled Win kin H flf ^ 15 ; in Tibetan mu-men, Mongol and Manchu 

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, kahrupdl, is transmitted. 10 This word 
corresponds to New Persian kahruba, a compound formed with kdh 
("straw") and ruba ("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. 

I This refers to Wood, Journey to the Oxus, p. 263. 
* See, further, M. Bauer, Precious Stones, p. 442. 

4 Ch. 22, p. 65. 

s The Diamond, p. 53. 

9 Ta T'an leu tien, Ch. 22, p. 8. 

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

8 In the Wei lio and Hou Han Su (cf. Chavannes, Toung Pao, 1907, p. 182). 

9 Nan Si, Ch. 79, p. 8; Wei Su, Ch. 102, p. 5 a; Sui Su, Ch. 83, p. 7 b. The Sui 
Su has altered the name hu-p'o into Sou-p'o §fc §%, in order to observe the tabu 
of the name Hu in Li Hu $ $£, the father of the founder of the T'ang dynasty. 
Amber (also coral and silver) is attributed to Mount Ni J§ tfj in the country Fu-lu^ni 
f^ Z§1 Tfe to the north of Persia, also to the country Hu-se-mi P£ $1 $? (Wei Su, 
Ch. 102, p. 6 b). 

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

II Analogies occur in all languages: Chinese Si-kiai $fe ^fc ("attracting mustard- 
seeds"); Sanskrit Irinagrdhin ("attracting straw"); Tibetan sbur len or sbur Ion, 
of the same meaning: French (obsolete) tire-paille. Another Persian word for amber 
is Sahbari. 

522 Sino-Iranica 

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 carabt; Byzantine Kepa(3t; Cumanian ckarabar. 
Under the Ming, amber is listed as a product of Herat, Khotan, and 
Samarkand. 2 A peculiar variety styled "gold amber" {kin p*o ^ IS) 
is assigned to Arabia (T'ien-fah). 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 Bul- 
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. 6 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- 

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

3 Ibid., Ch. 91, p. 20. 

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

5 Leclerc, Traits 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, 11, § 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 Yuh-c'ah in Yun-nan, and even on the sacred Hwa-san in 


G. Jacob 3 has called attention to the fact that the supposition of a 
derivation of the Chinese word from Pahlavi kahrupdi is confronted 
with unsurmountable difficulties of a chronological character. The 
phonetic difficulties are still more aggravating; for Chinese hu-p'o S£ 18 
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'uh. 6 

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 xuba 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 H, Ch. 78, p. 7. 

2 Cf. Hwa yoti^fc^ jg, Ch. 3, p. 1 (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 dprAfw (In Syria quoque feminas verticillos 
inde facere et vocare harpaga, quia folia paleasque et vestium fimbrias rapiat. — 
Pliny, xxxvii, II, § 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 Sino-Iranica 

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. Schoef 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 Zi §k tffl !& 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 most 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 Kuh 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. 
* Ibid., p. 44. 
8 Ibid., p. 59. 

4 Ibid., p. 246. 

5 XXXII, 11. 

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 l'Extreme-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, nor do we 
know an ancient Iranian name for coral. Solinus informs 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, ffl 3$ San-hit, *san-gu (Japanese 
san-go), possibly is of foreign origin, but possibly it is not/' 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 fanod or translates nerkupa. The common word in New 
Persian is marjan (hence Russian marZan)', other designations are 
birbal, xuruhak or xurohak, bussad or bissad (Arabic bessed or bussad). 
In Armenian it is bust. 7 

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

1 Cou Su, Ch. 50, p. 6; Sui Su, Ch. 83, p. 7 b; regarding coral in Fu-lu-ni, see 
above, p. 521 , note 9. 

2 T'an Su, Ch. 221 B, p. 6 b. The Lian Su (Ch. 54, p. 14 b) attributes to Persia 
coral-trees one or two feet high. 

* Ts'ien Han Su, Ch. 96 A, p. 5. This passage (not Hou Han Su, Ch. 118, as stated 
by Hirth, Chau Ju-kua, p. 226, after Bretschneider) contains the earliest mention 
of the word San-hu. 

* 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. 

8 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 pddazahar (see HUbschmann, Persische Studien, p. 193). Steingass 
gives also pdnzahr. The derivation from bad "wind" (H. FUhner, Janus, Vol. VI, 
1 90 1, p. 317) is not correct. 

9 Lander des Islam, p. 45. 

526 Sino-Iranica 

by Pelliot. 1 Pelliot, however, noticed well that what the Chinese 
describe as p'o-so or mo-so m l£ 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 pazakr and a Persian bezoar, the Chinese would not for 
a moment fail to inform us that p'o-so is the Pose niu-hwah or Persian 
bezoar; but they say nothing to this effect. On the contrary, the texts 
cited under this heading in the Pen ts x 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-cen points to the Ken sin yii ts'e 
JfE ^r 3£ Wf, written about 1430, as saying that the stone comes from 
San-fu-ts'i (Palembang on Sumatra). 6 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 x o-so stone is of animal origin, but, on the con- 
trary, they state explicitly that it is of mineral origin. The Ken sin yii 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 7 refers only to a stone of 
mineral origin. 

1 T'oung Pao, 1912, p. 438. 

2 The initial of the Persian word would require a labial surd in Chinese. Whether 
the p'o-sa §| g|| 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 H pen ts'ao 
of the third century, and by T'ao Huh-kin. 

4 Ch. 10, p. 10 b. 

5 This text is cited in the same manner in the Tun si yan k'ao of 161 8 (Ch. 3, 
p. 10). Cf. F. de Mely, 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'in ts'iian San kil ts'ao mu 
kitf&lUJg^fctZby Li Te-yu ^£§l& (a.d. 787-849) as a curious 
stone preserved in the P'o-so Pavilion south of the C'ah-tien it WL in 

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-Sen 1 has devoted a separate article to it under the name t^ -§• Za-ta, 
and has recognized it as a kind of bezoar; in fact, it follows immediately 
his article on the Chinese bezoar (niu-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 Cifah wai ki, 3 and published in 1623, mention an animal 
of Borneo resembling a sheep and a deer, called pa-tsaW ffi ft HI, 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 guliga; 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. 

2 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 
poo i h ffl X % 51 5& % by Wan 2en-yu ]£ fc M of the T'ang (p. 20 b). 
Cf. also the Sii K'ien $u |ff jjfr ^, written by Can Cu 3H fff in 1805 (Ch. 6, p. 8, 
ed. of Ytie ya Van ts'un Su). The Yakut know this stone as sata (Boehtlingk, Jakut. 
Worterbuch, 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); 
Vambery, Primitive Cultur, p. 249; Potanin, Tangutsko-Tibetskaya Okraina 
Kitaya, Vol. II, p. 352, where further literature is cited. 

3 Ch. 1, p. 11 (see above, p. 433). 

4 This form comes very near to the pajar of Barbosa in 15 16. 

6 Cf. the Lu Ian kun H 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 Sino-Iranica 

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


92. H St sa-pao, *sa5(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 
xfadra-pdvan (xSgpava, xSagapdvd) , which resulted in Assyrian axSadar- 
apdn or axladrapdn, Hebrew axaSdarfnim, 2 Greek aarpairris (Armenian 
Sahapand, Sanskrit ksatrapa). The Middle-Persian form from which the 
Chinese transcription was very exactly made must have been *sa0-pS,v 
or *xsa0-pav. The character sa renders also Middle and New Persian 
sar ("head, chief"). 3 

93. M H ^0 K'u-sa-ho, *Ku-sa5(r)-7wa, was the title rf 8 of the 
kings of Parsa (Persia). 4 This transcription appears to be based on an 
Iranian xSadva or xSarva, corresponding to Old Iranian *x§ayavan-, 
*xsaivan, Sogdian xSevan (" king ") . 5 It is notable that the initial spirant 
x is plainly and aptly expressed in Chinese by the element k'u, 6 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'Ecolefrangaise, Vol. Ill, pp. 665-671. 

a H. Pognon, Journal asiatique, 191 7, I, p. 395. 

3 R. Gauthiot, Journal asiatique, 191 1, II, p. 60. 

1 Sui Su, Ch. 83, p. 7 b. 

6 R. Gauthiot, Essai sur le vocalisme du sogdien, p. 97. See also the note of 
Andreas in A. Christensen, 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 Manichaean transcriptions it is expressed by P^ *xu ihu) ; see Cha- 
vannes and Pelliot, Traits manicheen, p. 25. 


530 Sino-Iranica 

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 a$x- for 
x$-, otherwise £ appears regularly save that $% takes the place of inter- 
vocalic xL 2 In view of our Sino-Iranian form, this rule should perhaps 
be reconsidered, but this must remain for the discussion of Iranian 

94. ^ !£? $a-ye, *sat(sa5)-ya. Title of the sons of the king of 
Persia (Wei iw, Ch. 102, p. 6; T'ai p K ih hwan yil ki, Ch. 185, p. 17). 
It corresponds to Avestan xlaBrya ("lord, ruler"). 3 The princes of 
the Sasanian empire were styled sa0ra5aran. 4 According to Sasanian 
custom, the sons of kings ruled provinces as "kings." 5 Regarding ix 
in transcriptions of Iranian names, cf. the name of the river Yaxartes 
HI i$t (Sui S» f Ch. 83, p. 4b) Yao-sa, that is *Yak-sa5(sar). As the 
Middle-Persian name is Xsart or Asart (Pazend A sard), 6 we are bound 
to assume that the prototype of the Chinese transcription was *Axsart 
or *Yaxsart. 

95. HI p ff i-tsan, but, as the fan-tsHe of the last character is indicated 
by ~Jf tfll, the proper reading is i-ts*at, *i-dza5, i-dza5, designation of the 
king of Parsa (^ A ^ or II £ W % Wei $u, Ch. 102, p. 6; Tai 
p'in hwan yil ki, Ch. 185, p. 17). The Chinese name apparently repre- 
sents a transcription of IxseS, the Ixsldh of al-Beruni, title of the 
kings of Sogd and Fergana, a dialectic form of Old Persian xlayaQiya? 
IxseS is the Avestan xSaeta ("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. 

2 Cf. also Gauthiot, op. cit., p. 54, § 6tv 

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

4 A. Christensen, op. cit., p. 20. Cf. Old Persian x$cm, xSacam ("royalty, 
kingdom"), Avestan xSaOrem, Sanskrit ksatram (A. Meillet, Grammaire du vieux 
perse, p. 143) ; xsaOrya corresponds to Sanskrit ksatriya. 

6 Noldeke, 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 (xsayadiya 
parsaiy, "king in Persia"); F. W. K. Muller, Ein Doppelblatt aus einem mani- 
chaischen Hymnenbuch, p. 31. 

Titles of 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. WHfc^- fan-pu-$wai, *pwan-bu-zwi5, designation of the queen 
of Parsa (Wei $u, Ch. 102, p. 6; T'ai p x ih hwan yii ki, Ch. 185, p. 17). 
The foundation of this transcription is presented by Middle Persian 
banbu$n, 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 wybyHh ("consort"). 5 

97. St #i *S mo-hu-t'an, *mak-ku(mag-gu)-dan. Officials of 
Persia in charge of the judicial department ^ S ft ffl. M (Wei $u, 
Ch. 102, p. 6). K. Hori 6 has overlooked the fact that the element 
Van 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 fH M *muk-gu (mu-hu) is based on New Persian muy, 
mdy. 7 The ending dan reminds one of such formations as herbeban 
("judge") and mobehan mobeh ("chief of the Magi"), the latter being 
Old Persian magupati, Armenian mogpet, Pahlavi ntaupat, 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 Extrfime-Orient, p. 11. 

2 Spiegel Memorial Volume, p. 248. 

3 Tabari, p. 452. 

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

6 R. Gauthiot, Essai sur le vocalisme du sogdien, pp. 59, 112. The three afore- 
mentioned titles had already been indicated by Abel-Remus at (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 

8 Spiegel Memorial Volume, p. 248. 

7 Chavannes and Pelliot, Traite" manichgen, p. 170. Accordingly this example 
cannot be invoked as proving that tnuk might transcribe also tnak, as formerly 
assumed by Pelliot {Bull, de I'Ecole francaise, Vol. IV, p. 312). 

8 Horn, Neupersische Etymologie, No. 984; and Hubschmann, Persische 
Studien, p. 123. 

532 Sino-Iranica 

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

98. tJ2 i£ ft ni-hu-han, *ni-hwut-7an. Officials of Persia who have 
charge of the Treasury (Wei £w, Ch. 102, p. 6). The word, in fact, is a 
family-name or title written by the Greek authors Naxopayav, Naxoepyav, 
Xapvaxopyavrjs (prefixed by the word sar, "head, upper"). FirdausI 
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 *Nixuryan or Nexuryan; and, indeed, the 
form Nixorakan is also found. 2 

99. iiil J^. Wi ti-pei-p'o, *di-pi-bwi5(bir, wir). Officials of Persia 
who have charge of official documents and all affairs (Cou i«, Ch. 50, 
p. 5b). In the parallel passage of the Wei iw (Ch. 102, p. 6), the second 
character is misprinted «F tsao, 3 *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 dipnr or dapir, New Persian dibir or datnr 
(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 diptr and diinr are contractions from diptvar. 
This word, as follows from the definition, appears to have comprised 
also what was understood by devdn, the administrative chanceries of 
the Sasanian empire. N 

100. t§l J$t ffl i& 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 iw, 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 Naxlrajan q and j represent 
Pahlavi g. The reconstructions attempted by Modi (Spiegel Memorial Volume, 
p. lix) of this and other Sino-Iranian words on the basis of the modern Chinese 
pronunciation do not call for any discussion. 

s 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. 

6 in, 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/Si^s), as observed by NOldeke, 1 should 
be read Argabides ('Apya(3i8r)s), the equivalent of ArgabeS. There 
is also a form apyaireTrjs 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 alkqfta, arkafta, or arkabta. 3 The above tran- 
scription is apparently based on the form *Argade ('Apyadrf) =Argabe5. 

101. I? $t %h 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 paikds {pat, "province"; kos, "guarding") 
9 # Jfc ^: Wei itt, Ch. 102, p. 6. The Cou £« (Ch. 50, p. 5b) 
has HI *sat, sar, in the place of the first character. The word corresponds 
to Middle Persian spahbed ("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; 4 the Chinese transcription, however, 
corresponds better to New Persian sipahbad, so that also a Middle- 
Persian form *spahba5 (-be5 or -bu5) may be inferred. 

102. 3l 5H nu-se-ta, *u-se-daS, used in the Chinese inscription dated 1489 
of the Jews of K'ai-fon fu in Ho-nan, in connection with the preceding name ??\\ Wi 
Lie-wei (Levi). 5 As justly recognized by G. Deveria, 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. 11 b) the Jews are styled Su-hu (Ju-hud) 

1 Tabari, p. 5. 

2 Christensen, op. cit., p. 27; N6ldeke, 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. 

8 J. Tobar, Inscriptions juives de K'ai-fong-fou, p. 44. 

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

534 Sino-Iranica 

}ft ^ or Cu-wu Ek 7C- This form can have been transcribed only on the basis of 
New Persian JuhG5 or JahQS 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 Yahut, 
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. 


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 Kiun JJS 1t ££^ and cloth of Su (Se-c'wan) 
§3 ^ . What this textile exactly was is not known. 2 Both these articles 
hailed from what is now Se-5'wan, Kiun being situated in Zun Sou Ufa. #H 
in the prefecture of Kia-tih, 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-c'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-c'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 Kiun, also called 
£!j, 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. 

2 Assuredly it was not silk, as arbitrarily inferred by F. 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 Ukaku-dake) is said to have been introduced into Japan from Liukiu. 
Forbes and Hemsley, Journal Linnean Soc, Vol. XXXVI, p. 443. 


536 Sino-Iranica 

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

The Pet hu lu 1 has the following notice on staves of the square 
bamboo: "C'eh cou §£ #H (in Kwah-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 K'iuh 3%, mentioned by Can K'ien. Such are produced also in Yun 
cou ra$ 'M, 2 the largest of these reaching several tens of feet in height. 
According to the Ceh hh tsi jE M ft, there are in the southern ter- 
ritory square bamboo staves on which the white cicadas chirp, and 
which C'en Cen-tsie W> M tp has extolled. Moreover, Hai-yen 9$ H 3 
produces rushes (lu M., Phragmites communis) capable of being made 
into staves for support. P'an cou M ^H 4 produces thousand-years ferns 
^ He W. and walking-sticks which are small and resemble the palmyra 
palm jt & (Borassus rflabelliformis) . There is, further, the su-tsie 
bamboo J8l W 1T, from which staves are abundantly made for the 
Buddhist and Taoist clergy, — all singular objects. According to the 
Hut tsui H" Jt, the t'un M bamboo from the Cen River $£ JH 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-c'wan, and also grows on the mountain Fei-lai-fuh 3$ 2fS Hr 
on the West Lake in Ce-kiah; the knots of this bamboo are prickly, 
hence it is styled in Se-6'wan tse lu JW 1T ("prickly bamboo"). 

According to the Min siao ki P5 /h IS, 6 written by Cou Lian-kun 
ffl $& X in the latter part of the seventeenth century, square bamboo 
and staves made from it are produced in the district of Yun-tih ^C % 
in the prefecture of T'ih-Cou and in the district of T'ai-nih ^ ^ 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-2ou, 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-mih hien, forming the pref ectural city of Kao-cou f u, Kwan-tun. 
6 Ch. 8, p. 9 (ed. of Si yin Man ts'un Su). 

6 Ed. of Swo tin, p. 17. 

7 The San hai kin mentions the "narrow bamboo (hia cu $$£ 1T) 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 Kiuh. According to the 
Kwan li, the Kiun bamboo occurred in the districts of Nan-kwan ff^J llf (at present 
Nan-k'i f|| $£) and Kiuh-tu in Se-5*wan. The Memoirs of Mount Lo-fou (Lo-fou 
San ki) in Kwan-tuh 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" ffi ^g 'fj*. These texts are cited in the T'ai p'in yii Ian (Ch. 963, p. 3). 

Irano-Sinica — The Square Bamboo, Silk 537 

It is said to occur also in the prefecture of Teh-cou $£ 4H, 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-6'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 Pet hu lu, 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 .^\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 llri) is fre- --'*"' 
quently mentioned by FirdausI as playing a prominent part in Persian 
decorations. 6 He also speaks of a very fine and decorated Chinese silk 
under the name parniyan, corresponding to Middle Persian parntkan? 
Iranian has a peculiar word for "silk," not yet satisfactorily explained: 
Pahlavi *apresum, *aparesum; New Persian abre*>urn, abreiam (Arme- 

1 San tun t'un li, Ch. 9, p. 6. 

2 See K'ien Su 3^ ilF, Ch. 4, p. 7 b (in Yue ya fan ts'un Su, t'ao 24) and Su K'ien 
hi, Ch. 7, p. 2 b (ibid.). Cf. also tu p'u sian lu 1t Wt f£ £&. written by Li K'an 
^ ffX "! I2 99 (Ch. 4, p. I b; ed. of Ci pu tsu lai 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. 

* Hirth, Chinesische Studien, p. 10. 

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

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

7 HtJBSCHMANN, Persische Studien, p. 242. 

538 Sino-Iranica 

nian, loan-word from Persian, aprUum)\ hence Arabic ibarisam or 
ibrlsam; Pamir dialects warSum, war$um, Sugni wrelom, etc.; Afghan 
writs' am. 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 Urgek 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 (pent- tire) have 
existed. This, however, was assuredly not the case. We know that the 
termination V jrcl, 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 V 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 V. 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-R6musat. 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. Etymologie, No. 65. 
The derivation from Sanskrit ksautna is surely wrong. Bulgar ibriUm, Rumanian 
ibrisin, are likewise connected with the Iranian series. 

2 Cf. Klaproth, Conjecture sur l'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 M£moires relatifs a TAsie, 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, 191 2, 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, 

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 lirgek: 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 Mk, Lithuanian szilkai), 
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, kamxd, kimxd (Arabic kimxaw, Hin- 
dustani kamxdb), designating a "gold brocade," as I formerly ex- 
plained, 1 may be derived from Chinese $S ^£ kin-hwa, *kim-xwa. 

3-4. Of fruits, the West is chiefly indebted to China for the peacm I 
(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 Persicaj 
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 
the first to cultivate this fruit from ancient times. Previous authors \ 
have justly connected the westward migration of peach and apricot I 
with the lively intercourse of China and western Asia following Ca6j 
K'ien's mission. 5 Persian has only descriptive names for these fruits, 
the peach being termed lajt-alu, ("large plum"), the apricot zard-alu 

1 Toting Pao, 1916, p. 477; Yule, Hobson-Jobson, p. 484. 

2 xv, 11, 13. 

' De Candolle (Origin of Cultivated Plants, p. 222) is mistaken in crediting 
Theophrastus with the knowledge of the peach. Joret (Plantes dans l'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 Sino-Iranica 

("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 
Hiian Tsah. 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 Cinabhukti ("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 clnani ("Chinese fruit"); 
and the pear, clnarajaputra ("crown-prince of China"). These names 
are still prevalent. 3 Although Hiian Tsan 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 Hiian Tsah'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 
leri, WaxI liwan or loan (but Sariqoll noS, 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 Kagmlrl (tser , tser-kul) ; further, in West-Tibetan lu-li or lo-li, 
Balti su-ri, Kanaurl lul (other Tibetan words for "apricot" are k'am-bu, a-$u, and 
la-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 tserduti, 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 T'an Si yil hi, Ch. 4, p. 5. 

s There are a few other Indian names of products formed with "China": 
clnapitfa ("minium"), clnaka ("Panicum miliaceum, fennel, a kind of camphor"), 
clnakarpura ("a kind of camphor"), clnavanga ("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. Watters, 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 

5. As indicated by the Persian name dar-llnl or dar-lln ("Chinese 
wood" or "bark"; Arabic dar ?lni), cinnamon was obtained by the 
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 
cinnamon production, and the tree {Cinnamomum zeylanicum) is a native 
of the Ceylon forests. 5 The bark of this tree is also called dar-cini. 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 
Sabasans, 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'antiquite\ Vol. II, p. 281. 

* Leclerc, Traite" des simples, Vol. II, pp. 68, 272. The loan-word darilenik 
in Armenian proves that the word was known in Middle Persian (*dar-i £gnik) ; cf . 
Hubschmann, Armen. Gram., p. 137. 

* G. Ferrand, Textes relatifs a l'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. 

8 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. 

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

10 xii, 42. 

542 Sino-Iranica 

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-Arnolt: 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-U [?]. The -t in the Egyptian represents the feminine suffix." As 
may be seen from O. Schrader, 7 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 Kaala is derived from Hebrew qest'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. 1 19-120. 

5 Thus also Fluckiger and Hanbury (Pharmacographia, p. 520), whose 
argumentation is not sound, as it lacks all sense of chronology. The Persian term 
dar-cinl, 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 

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 ££ (*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, in 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 Kwah-si, Kwah- 
tuh, 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 Kwah-tuh, 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 
t£ $£) is first mentioned by T'ao Huh-kih (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 
jat the time of Herodotus or earlier. The earliest date we may assume 
£or any navigation from the coasts of Indo-China into the Indian Ocean 
\s 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 F6e 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. 

2 Cf. Pelliot, Toung Pao, 1912, pp. 457-461. 

1 The Malabar cinnamon is mentioned by Marco Polo (Yule's ed., Vol. II, 
p. 389) and others. 

544 Sino-Iranica 

or the attribution of certain products to China, is not always to be 
understood literally. Sometimes it merely refers to a far-eastern 
product, sometimes even to an Indian product, 1 and sometimes to 
products handled and traded by the Chinese, regardless of their pro- 
venience. Such cases, however, are exceptions. As a rule, these Persian- 
Arabic terms apply to actual products of China. 

Schlimmer 2 mentions under the name Killingea monocephala the 
zedoary of China: according to Piddington's Index Plantarum, it should 
be the plant furnishing the famous root known in Persia as jadware 
xitai ("Chinese jadvar"); genuine specimens are regarded as a divine 
panacea, and often paid at the fourfold price of fine gold. The identifica- 
tion, however, is hardly correct, for K. monocephala is kin niu ts'ao 
^ 4 1 ^ in Chinese, 3 which hardly holds an important place in the 
Chinese pharmacopoeia. The plant which Schlimmer had in mind 
doubtless is Curcuma zedoaria, a native of Bengal and perhaps of China 
and various other parts of Asia. 4 It is called in Sanskrit nirvisa ("poison- 
less") or sida, in KuSa or Tokharian B viralom or wiralom, 5 Persian jad- 
var, Arabic zadvar (hence our zedoary, French zedoaire). Abu Mansur 
describes it as zarvar, calling it an Indian remedy similar to Costus and 
a good antidote. 6 In the middle ages it was a much-desired article of 
trade bought by European merchants in the Levant, where it was sold 
as a product of the farthest east. 7 Persian zarumbdd, Arabic zeronbdd, 
designating an aromatic root similar to zedoary, resulted in our zer- 
umbet. s While it is not certain that Curcuma zedoaria occurs in China 
(a Chinese name is not known to me), it is noteworthy that the Persians, 
as indicated above, ascribe to the root a Chinese origin: thus also 
ka%ur (from Sanskrit karcura) is explained in the Persian Dictionary of 

1 Such an example I have given in T'&ung Pao, 1915, p. 319: biS, an edible 
aconite, does not occur in China, as stated by Damlrl, but in India. In regard to 
cubebs, however, Garcia da Orta (C. Markham, Colloquies, p. 169) was mis- 
taken in denying that they were grown in China, and in asserting that they are 
called kabab-£inl only because they are brought by the Chinese. As I have 
shown {ibid., pp. 282-288), cubebs were cultivated in China from the Sung period 

2 Terminologie, p. 335. 

3 Also this identification is doubtful (Stuart, Chinese Materia Medica, 
p. 228). 

4 W. Roxburgh, Flora Indica, p. 8; Watt, Commercial Products of India, 
p. 444, and Dictionary, Vol. II, p. 669. 

B S. Levi, Journal asiatique, 191 1, II, pp. 123, 138. 

6 Achundow, Abu Mansur, p. 79. See also Leclerc, Traits des simples, Vol. I, 
P- 347- 

7 W. Heyd, Histoire du commerce du levant, Vol. II, p. 676. 

8 Yule, Hobson-Jobson, p. 979. 

Irano-Sinica — Zedoary, Ginger 545 

Steingass as "zedoary, a Chinese root." Further, we read under tndh- 
parwdr or parwtn, "zedoary, a Chinese root like ginger, but perfumed." 

7. Abu Mansur distinguishes under the Arabic name zanjabil three 
kinds of ginger (product of Amomum zingiber, or Zingiber officinale), — 
Chinese, Zanzibar, and Melinawi or Zurunbaj, the best being the 
Chinese. 1 According to Steingass, 2 Persian anqala denotes "a kind 
of China ginger." 3 The Persian word (likewise in Arabic) demonstrates 
that the product was received from India: compare Prakrit singabera, 
Sanskrit qrhgavera (of recent origin), 4 Old Arabic zangabil, Pahlavi 
langawr, New Persian tankalil, Arabic-Persian zanjabil, Armenian 
snrvel or snkrvil (from *singivel), Greek {iyyifiepis, Latin zingiberi; 
Madagasy iakaviru (Indian loan-word). 5 

The word galangal, denoting the aromatic rhizome of Alpinia 
galanga, is not of Chinese origin, as first supposed by D. Hanbury, 6 
and after him by Hirth 7 and Giles. 8 The error was mainly provoked 
by the fact that the Arabic word from which the European name is 
derived was wrongly written by Hanbury khalanjan, while in fact it is 
khulanjan {xulandZan), Persian xawalinjdn. The fact that Ibn Khor- 
dadzbeh, who wrote about a.d. 844-848, mentions khulanjan as one of 
the products of China, 9 does not prove that the Arabs received this 
word from China; for this rhizome is not a product peculiar to China, 
but is intensively grown in India, and there the Arabs made the first 
acquaintance of it. Ibn al-Baitar 10 states expressly that khulanjan 
comes from India; and, as was recognized long ago, the Arabic word 
is derived from Sanskrit kulanja, 11 which denotes Alpinia galanga. 
The European forms with ng (galangan, galgan, etc.) were suggested by 
the older Arabic pronunciation khulangan. 12 In Middle Greek we have 

1 Achundow, Abu Mansur, p. 76. 

2 Persian Dictionary, p. 113. 

I Concerning ginger among the Arabs, cf. Leclerc, Traite* des simples, Vol. II, 
p. 217; and regarding its preparation, see G. Ferrand, Textes relatifs a l'Extreme- 
Orient, p. 609. 

4 Cf. the discussion of E. Hultzsch and F. W. Thomas in Journal Roy. As. Soc, 
1912, pp. 475, 1093. See also Yule, Hobson-Jobson, p. 374. 

6 The curious word for "ginger" in Kuca or Tokharian B, tvankaro (S. Levi, 
Journal asiatique, 191 1, II, pp. 124, 137), is not yet explained. 

6 Science Papers, p. 373. 

7 Chinesische Studien, p. 219. 

8 Glossary of Reference, p. 102. 

9 G. Ferrand, Textes relatifs a l'Extr6me-Orient, p. 31. 

10 Ibid., p. 259. Cf. also Achundow, Abu Mansur, p. 60. 

II Roediger and Pott, Z. K. d. Morgenl., Vol. VII, 1850, p. 128. 

n E. Wiedemann (Sitzber. Phys.-Med. Soz. Erl., Vol. XLV, 1913, p. 44) gives 
as Arabic forms also xaulangdd and xalangan. 

546 Sino-Iranica 

KokovT^ia, xauXtfei', and 7aXaYyd; in Russian, kalgdn. The whole group 
has nothing to do with Chinese kao-liah-kiah. 1 Moreover, the latter 
refers to a different species, Alpinia officinarum; while Alpinia galanga 
does not occur in China, but is a native of Bengal, Assam, Burma, 
Ceylon, and the Konkan. Garcia da Orta was already well posted on 
the differences between the two. 2 

8. Abu Mansur mentions the medical properties of mdmirdn. 3 
According to Achundow, 4 a rhizome originating from China, and 
called in Turkistan momiran, is described by Dragendorff, and is re- 
garded by him as identical with the so-called mishmee (from Coptis 
teeta Wall.), which is said to be styled mamiraUn in the Caucasus. He 
further correlates the same drug with Ranunculus ficaria {x&'&bviov 
to fxiKpov), subsequently described by the Arabs under the name 
mamirun. Al-Janki is quoted by Ibn al-Baitar as saying that the 
mdmirdn comes from China, and that its properties come near to 
those of Curcuma; 1 ' these roots, however, are also a product of Spain, 
the Berber country, and Greece. 6 The Sheikh Daud says that the best 
which comes from India is blackish, while that of China is yellowish. 
Ibn Batuta 7 mentions the importation of mdmirdn from China, saying 
that it has the same properties as kurkum. Hajji Mahomed, in his 
account of Cathay (ca. 1550), speaks of a little root growing in the 
mountains of Succuir (Su-cou in Kan-su), where the rhubarb grows, 
and which they call Mambroni Cini (mdmlrdn-i Clnl, "mamiran of 
China"). "This is extremely dear, and is used in most of their ail- 
ments, but especially where the eyes are affected. They grind it on 
a stone with rose-water, and anoint the eyes with it. The result is 
wonderfully beneficial." 8 In 1583 Leonhart Rauwolf 9 mentions 

1 Needless to say that the vivisections of Hirth, who did not know the Sanskrit 
term, lack philological method. \ 

2 Markham, Colloquies, p. 208. Garcia gives lavandou as the name used in 
China; this is apparently a corrupted Malayan form (cf. Javanese laos). In Java, he 
says, there is another larger kind, called lancuaz; in India both are styled lancuaz. This 
is Malayan lenkuwas, Makasar lankuwasa, Cam lakuah or lakuak, Tagalog lankuas. 
The Arabic names are written by Garcia calvegiam, chamligiam, and galungem; the 
author's Portuguese spelling, of course, must be taken into consideration. 

3 Achundow, Abu Mansur, p. 138. 

4 Ibid., p. 268. 

5 Leclerc, Traite" des simples, Vol. II, p. 441. Dioscorides remarks that the 
sap of this plant has the color of saffron. 

6 In Byzantine Greek it is ixa/i-qpe or nempkv, derived from the Persian- Arabic 

7 Ed. of Defremery and Sanguinetti, Vol. II, p. 186. 

8 Yule, Cathay, new ed., Vol. I, p. 292. 

9 Beschreibung der Raiss inn die Morgenlander, p. 126. 

Irano-Sinica — Mamiran, Rhubarb 547 

the drug mamirani tchini for eye-diseases, being yellowish like Curcuma. 

Bernier mentions mamiran as one of the products brought' by the 
caravans from Tibet. Also according to a modern Mohammedan source, 
mamiran and rhubarb are exported from Tibet. 1 

Mamira is a reputed drug for eye-diseases, applied to bitter roots 
of kindred properties but of different origin. By some it is regarded as 
the rhizome of Coptis teeta (ttta being the name of the drug in the Mishmi 
country) ; by others, from Thalictrum joliosum, a tall plant common 
throughout the temperate Himalaya and in the Kasia Hills. 2 In another 
passage, however, Yule 3 suggests that this root might be the ginseng 
of the Chinese, which is highly improbable. 

It is most likely that by mamira is understood in general the root of 
Coptis teeta. This is a ranunculaceous plant, and the root has some- 
times the appearance of a bird's claw. It is shipped in large quantities 
from China (Chinese hwan-lien !ic *H) via Singapore to India. The 
Chinese regard it as a panacea for a great many ills; among others, for 
clearing inflamed eyes. 

9. Abu Mansur discriminates between two kinds of rhubarb, — the 
Chinese (rlwand-i stni) and that of Khorasan, adding that the former 
is most employed. 4 Accordingly a species of rhubarb (probably Rheum 
ribes) must have been indigenous to Persia. Yaqut says that the finest 
kind grew in the soil of Nlsapur. 5 According to E. Boissier, 6 Rheum 
ribes occurs near Van and in Agerowdagh in Armenia, on Mount Pir 
Omar Gudrun in Kurdistan, in the Daena Mountain of eastern Persia, 
near Persepolis, in the province Aderbeijan in northern Persia, and in 
the mountains of Baluchistan. There is a general Iranian name for 
"rhubarb": Middle Persian rewas, New Persian rewas, rewand, riwand 
(hence Armenian erevant), Kurd riwas, rlbas; Baluci ravaS; Afghan 
rawa!!;. , 7 The Persian name has penetrated in the same form into Arabic 

1 Ch. Schefer, Histoire de l'Asie centrale par Mir Abdoul Kerim Boukhary, 
p. 239. Cf. also R. Dozy, Supplement aux dictionnaires arabes, Vol. II, p. 565. 

2 Yule, Hobson-Jobson, p. 548. 

* Cathay, Vol. I, p. 292. 

4 Achundow, Abu Mansur, p. 74. Chinese rhubarb is also called simply lini 
("Chinese") in Persian, sini in Arabic. 

8 Barbier de Meynard, Diet. geogr. de la Perse, p. 579. 

• Flora Orientalis, Vol. IV, p. 1004. Rheum ribes does not occur in China or 
Central Asia. 

7 The Afghan word in particular refers to Rheum spiciforme, which grows wild 
and abundantly in many parts of Afghanistan. When green, the leaf-stalks are 
called rawds; and when blanched by heaping up stones and gravel around them, 
lukri; when fresh, they are eaten either raw or cooked (Watt, Dictionary, Vol.VI, 
p. 487). The species under notice occurs also in Kan-su, China: Forbes and 

548 Sino-Iranica 

and Turkish, likewise into Russian as reven' and into Serbian as reved. 
It is assumed also that Greek frrjov (from *rewon) and pa. are derived from 
Iranian, and it is more than likely that Iran furnished the rhubarb 
known to the ancients. The two Greek names first appear in Dios- 
corides, 1 who states that the plant grows in the regions beyond the 
Bosporus, for which reason it was subsequently styled rha ponticum 
or rha barbarum (hence our rhubarb, Spanish ruibarbo, Italian rabarbaro, 
French rhubarbe), — an interesting case analogous to that of the Hu 
plants of the Chinese. In the fourth century, Ammianus Marcellinus 2 
states that the plant receives its name from the River Rha ('Pa, Finnish 
Rau, Rawa), on the banks of which it grows. This is the Volga, but the 
plant does not occur there. It is clear that Ammianus' opinion is 
erroneous, being merely elicited by the homophony of the names of 
the plant and the river. Pliny 3 describes a root termed rhacoma, which 
when pounded yields a color like that of wine but inclining to saffron, 
and which was brought from beyond the Pontus. Certain it is that 
this drug represents some species of Rheum, in my opinion identical 
with that of Iran. 4 There is no reason to speculate, as has been done by 
some authors, that the rhubarb of the ancients came from China; for 
the Chinese did not know rhubarb, as formerly assumed, from time 
immemorial. This is shown at the outset by the composite name ta 
hwan 3s. if ("the great yellow one") or hwan liah lit Ji("the yellow 
good one"), merely descriptive attributes, while for all genuinely ancient 
plants there is a root-word of a single syllable. The alleged mention of 
rhubarb in the Pen kin or Pen ts x ao, attributed to the mythical Emperor 
Sen-nun, proves nothing; that work is entirely spurious, and the text 
in which we have it at present is a reconstruction based on quotations 
in the preserved Pen-ts'ao literature, and teems with interpolations and 
anachronisms. 6 All that is certain is that rhubarb was known to the 

Hemsley, Journal Linnean Soc, Vol. XXVI, p. 355. There is accordingly no rea- 
son to seek for an outside origin of the Iranian word (cf. Schrader, Reallexikon, 
p. 685). The Iranian word originally designated an indigenous Iranian species, 
and was applied to Rheum officinale and palmatum from the tenth century onward, 
when the roots of these species were imported from China. 

1 in, 2. Theophrastus is not acquainted with this genus. 

J XXII. vin, 28. 

* xxvii, 105. 

4 Fltjckiger and Hanbury (Pharmacographia, p. 493) state, "Whether pro- 
duced in the regions of the Euxine (Pontus), or merely received thence from remoter 
countries, is a question that cannot be solved." The authors are not acquainted 
with the Iranian species, and their scepticism is not justified. 

6 It is suspicious that, according to Wu P'u of the third century, Sen Nun and 
Lei Kuh ascribed poisonous properties to ta hwan, while this in fact is not true. 
The Pen kin (according to others, the Pie lu) states that it is non-poisonous. 

Irano-Sinica — Rhubarb 549 

Chinese in the age of the Han, for the name ta hwan occurs on one of 
the wooden tablets of that period discovered in Turkistan by Sir A. 
Stein and deciphered by Chavannes. 1 

Abu Mansur, as cited above, is the first Persian author who speaks 
of Chinese rhubarb. He is followed by a number of Arabic writers. 
It is therefore reasonable to infer that only in the course of the tenth 
century did rhubarb develop into an article of trade from China to 
western Asia. In n 54 EdrisI mentions rhubarb as a product of China 
growing in the mountains of Buthink (perhaps north-eastern Tibet). 2 
Ibn Sa'id, who wrote in the thirteenth century, speaks of the abundance 
of rhubarb in China. 3 Ibn al-Baitar treats at great length of rawend, 
by which he understands Persian and Chinese rhubarb, 4 and of ribds, 
"very common in Syria and the northern countries," identified by 
Leclerc with Rheum ribes} 

Marco Polo relates that rhubarb is found in great abundance over 
all mountains of the province of Sukchur (Su-cou in Kan-su), and that 
merchants go there to buy it, and carry it thence all over the 
world. 6 In another passage he attributes rhubarb also to the mountains 
around the city of Su-cou in Kian-su, 7 which, Yule says, is believed by 
the most competent authorities to be quite erroneous. True it is that 
rhubarb has never been found in that province or anywhere in middle 
China; neither is there an allusion to this in Chinese accounts, which 
restrict the area of the plant to Sen-si, Kan-su, Se-c'wan, and Tibet. 
Nevertheless it would not be impossible that at Polo's time a sporadic 
attempt was made to cultivate rhubarb in the environs of Su-cou. Friar 
Odoric mentions rhubarb for the province Kansan (Kan-su), growing 
in such abundance that you may load an ass with it for less than six 
groats. 8 

Chinese records tell us very little about the export-trade in this 
article. Cao 2u-kwa alone mentions rhubarb among the imports of 

1 Documents chinois d^couverts dans les sables du Turkestan oriental, p. 115, 
No. 527. 

5 W. Heyd, Histoire du commerce du levant, Vol. II, p. 665. See also Fluckiger 
and Hanbury, Pharmacographia, pp. 493-494. 

1 G. Ferrand, Textes relatifs a 1'ExtrSme-Orient, p. 350. 

4 Leclerc, Traits des simples, Vol. II, pp. 155-164. 

5 Ibid., p. 190. This passage was unknown to me when I identified above the 
Persian term riwand with this species, arriving at this conclusion simply by consult- 
ing Boissier's Flora. 

6 Yule, Marco Polo, Vol. I, p. 217. 

7 Ibid., Vol. II, p. 181. 

8 Yule, Cathay, Vol. II, p. 247. 

550 Sino-Iranica 

San-fu-ts'i (Palembang) and Malabar. 1 In vain also should we look in 
Chinese books for anything on the subject that would correspond to the 
importance attached to it in the West. 

Garcia da Orta (1562) held it for certain that "all the rhubarb 
that comes from Ormuz to India first comes from China to Ormuz by 
the province of Uzbeg which is part of Tartary. The fame is that it 
comes from China by land, but some say that it grows in the same 
province, at a city called £amarcander (Samarkand) . 2 But this is very 
bad and of little weight. Horses are purged with it in Persia, and I 
have also seen it so used in Balagate. It seems to me that this is the 
rhubarb which in Europe we called ravam turquino, not because it is 
of Turkey but from there." He emphasizes the point that there is no 
other rhubarb than that from China, and that the rhubarb coming to 
Persia or Uzbeg goes thence to Venice and to Spain; some goes to 
Venice by way of Alexandria, a good deal by Aleppo and Syrian Tripoli, 
all these routes being partly by sea, but chiefly by land; 3 the rhubarb 
is not so much powdered, for it is more rubbed in a month at sea than in 
a year going by land. 4 As early as the thirteenth century at least, as we 
see from Ibn al-Baitar, what was known to the Arabs as "rhubarb of 
the Turks or the Persians," in fact hailed from China. In the same 
manner, it was at a later time that in Europe "Russian, Turkey, and 
China rhubarb" were distinguished, these names being merely in- 
dicative of the various routes by which the drug was conveyed to 
Europe from China. 5 Also Christoval Acosta notes the corruption 
of rhubarb at sea and its overland transportation to Persia, Arabia, 
and Alexandria. 3 

1 Hirth, Chau Ju-kua, pp. 61, 88. 

2 Probably Rheum ribes, mentioned above. 

s Leonhart Rauwolf (Beschreibung der Raiss inn die Morgenlander, 1583, 
p. 461) reports that large quantities of rhubarb are shipped from India to Aleppo 
both by sea and by land. 

4 Cf . Markham, Colloquies,, pp. 390-392. 

5 In regard to the Russian trade in rhubarb see G. Cahen, Le livre de comptes 
de la caravane russe a P£kin, p. 108 (Paris, 191 1). 

6 Reobarbaro (medicina sing