A "Periplus" of Magan and Melua Author(s): John Hansman Source: Bulletin of the School of Oriental and African Studies,

University of London, Vol. 36, No. 3 (1973), pp. 554-587 Published by: Cambridge University Press on behalf of School of Oriental and African Studies Stable URL: http://www.jstor.org/stable/613582 . Accessed: 05/10/2011 23:56
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Cambridge University Press and School of Oriental and African Studies are collaborating with JSTOR to digitize, preserve and extend access to Bulletin of the School of Oriental and African Studies, University of London.

http://www.jstor.org

A PERIPLUS OF MAGAN

AND MELUHHA

By JOHNHANSMAN
(PLATE I)

Perhaps no question of historical-geographicalinterest relating to the ancient Middle East has been so frequently a subject of inquiry or brought a greater difference of opinion than has that of the location of the countries of Magan and Meluhha.1 The problems of identification are particularly confused because widely separate areas appear to be indicated at different periods. Old Babylonian and Sumerian references seem to place these regions at the lower end of the Persian Gulf and along the coast of the Indian Ocean beyond, while texts of the late Assyrian period indicate Egypt and Nubia/Ethiopia. To reconcile these textual inconsistencies Landsberger many years ago proposed a location for the earlier Magan in Arabia near the lower Persian Gulf and for earlier Meluhha, an unspecified region east of this Magan.2 He suggested that these two names were subsequently applied to Egypt and to Ethiopia because in the later Assyrian period some of the same materials (gold and ivory) were imported into Mesopotamia from the 'Southern Meluhha' as had formerly been supplied by the 'Eastern Meluhha '.3 Herzfeld 4 and Eilers 5 have proposed that Magan is to be found on both sides of the Gulf of Oman and both writers submit that the name Magan survives in Makran, a geographical designation for the coastal region of south-eastern Iran and southwestern Pakistan. Herzfeld further suggests that by Meluhha is meant the Red Sea coast of Egypt.6 Gelb identifies Magan with the southern shore of Arabia on the Persian Gulf, extending east from ancient Sumer up to and including Oman. Meluhha is the north shore of the Gulf including Iran past Elam and Anshan and also all the country east of there up to and including the Indus valley.7 Leemans favours the possibility of locating Meluhha in Sind and along the coast of western India.8 He follows Weidner9 in proposing Oman in southeastern Arabia for Magan.1o Kramer claims that most cuneiformists would agree that from about 1500 B.C.Magan and Meluhha corresponded roughly to ancient Egypt and Ethiopia. He argues, however, against the possibility of the name-change theory expressed by Landsberger. Kramer states that 'the verification of an hypothesis involving a name shift in the cuneiform documents for countries so important as Magan and Meluhha should be based on evidence
1 For abbreviations, see p. 584. 2 B. Landsberger, ZA, xxxv, 3, 1924, 217. 3ibid., p. 217, n. 2. 4E. Herzfeld, The Persian Empire, Wiesbaden, 1968, 63. 5Note by W. Eilers in W. Leemans, JESHO, III, 1, 1960, 29. 6 Herzfeld, op. cit., 63-4. 7I. J. Gelb, RA, LxIv, 1, 1970, 5. 8 W. F. Leemans, Foreign trade in the Old Babylonian period, Leiden, 1960, 162, 164. SE. F. Weidner, A f. O, xvI, 1, 1952, 22. o10 Leemans, Foreign trade, 159.

Z
C/I

I 1 c
3 tZ~ ijj

np~
8 d ~c~5~~ ar ~b ~o 9, 3 1 E:9)
v

~3,t b/ 3
v,3?

d
r r

c
?; ~

C) O
3~:=`

C
I*(e

?-

~= " hZ C =;~,~ ??
~ 4=--

fP Z?
C, e, cr e, ~ r\~ .~r~'h

)I

=~

-?

~b:

;B

c 4, ,cC3
Z s CI

g v,
~

3?

c r ~?S , ~ ~?

"

a~C) Flrlj w Q C--'

C~ r
7t

;i;i g t. iS: ;ir

pr; -~ Z~ul

*1~11~7N
??? s.. ??r?

d t~' ?,?, ........ .? ?~ ? ?? ?

?? .. ;,r

???

Zrtr?r.???r f? ~

4~
c"

?
.?? ~ r ~?=??? ?? ?? ?? ? ?? ?)?? ?? :f:j

Ir

?. ?.

??

Cr 4
4 c=

1?

??

??c
?? ...?,;:? ?

?? ?

o
rSI,

t)
--e

L ;I n - c. n,4
p ~Z

Cr~
71 r"

si"
5: ~~r ?? Zu=
~Fsr,~3~3?~V 'CI n ???? ?, '?.

(/??CYIA~c~,?ZeZCjljCD. ~rrr

ZZ

C~j BAtDC 33

t;

-c, cr
cr 3rcC

CUS24N
cc
?,

f
71

?~cc.

, z :? ?? rCZI ?

I
n jP S~ '?ir 7C

?.
f t ~

L "'3m
V) cr OCT o s3

r
Cu?

a
-j:

ur
B~ql:DC~s~4n, 'P ?,

4
a

Ir?.C'r
?. 5t

~5:

Z
r r?., U) I) U) ~

rr ,?~?r')e II Ifrtilir~tlc

~?1111111111~1~\\\\\II \t r.,~CD

,Illlff rtirttt, Ir"t

A PERIPLUS OF MAGAN AND MELUHITA

555

that is reasonably assured and decisive.' n He implies that the names were always to be associated with Egypt and Ethiopia and not transferred from areas located to the east of Mesopotamia. The present writer will propose that linguistically Magan has always been linked with the toponym Makran in what is now southern Baluchistan in Iran (fig. 1). Further, Meluhha is to be largely identified with the former Khanat of Kalat which includes most of Baluchistan in Pakistan and in the south Makran in Pakistan (fig. 1). In the geographical sense three duplicate late Assyrian inscriptions do indicate an equation of Magan with Egypt and numerous texts seem to associate Meluhha with Nubia and/or with the Sudan. However, a close study of the relevant passages shows that some of these later texts also use forms of the indigenous names of these African regions, hereafter to be referredto in the present context as the southern Magan and Meluhhla. This and other findings considered below suggest that Magan and Meluhla as applied to southern regions were toponyms perhaps borrowed from texts based on older Mesopotamianinscriptions concerned with eastern districts identified with these names. The later associations then may have arisen from logical but incorrect interpretations of the earlier records. To understand better the reasons for the considerable differences of opinion which exist on these questions and to establish a basis for objective discussion, it will be useful to list chronologically the textual references to Magan and Meluhha. Because Telmun or Dilmun (Bahrain) 12 is closely linked at various periods with the trade of both regions, and as this connexion will enable us to follow the continuation of the trade after references to Magan and Meluhha no longer occur, some passages relating to Telmun are included. In this first section only texts of the Old Akkadian and Sumerian periods or those originating from these periods are given. Wherever possible the name of the king with whom each text is associated is also given. Dynasty of Akkad Sargon (2370-2316 B.c.) caused the ships of Telmun, Meluhha, and Magan to be moored at the wharf before Agade.13 Manishtusu (2306-2292 B.c.) crossed the lower sea (the Persian Gulf) in ships. He defeated 32 kings and their cities on the other side of the sea and overthrew the whole country as far as the silver mines. He took stones from the mountains below the lower sea and had statues to be made from these.14 Naram-Sin (2291-2255 B.c.) subjugated Magan and killed Manui, Lord of Magan. In their mountains he quarried stone which was transported to Agade to make statues.'"
S. N. Kramer, An., xxxvII, 146, 1963, p. 114, n. 11. Telmun or Dilmun with Bahrain is set out in B. Cornwall, BASOR, 103, 1946, 3 f., and in B. Cornwall, JCS, vi, 4, 1952, 137 f. 13 G. A. Barton, The royal inscriptions of Sumer and Akkad, New Haven, 1929, 108. 14 ibid., 128-30. 15 ibid., 142.
11

12 The case for identifying

556

JOHN HANSMAN

Dynasty of Lagash Gudea (c. 2220 B.c.) brought naaESI (probably diorite) from the mountains of Magan to Lagash, where it was used to make his statue. He brought uu wood and gold dust from Melulha to Lagash. Brought also from Meluhha a kind of stone, probably carnelian." was GUG.GI.RIN.E, Third Dynasty of Ur Ibbi-Sin (2029-2006 B.c.), garments and wool were brought from store to trade for copper from Magan. Other imports to Ur at this period were Magan onions and Magan reeds.'7 From Meluhha came copper, A.AB.BA wood used for making chairs and dagger sheaths, mesu wood, and multicoloured birds of ivory.1s Assyria Ilu.uma (2004-1977 B.c.), in a year-formula text a throne is stated to have been made from AB.BA Meluhha wood.'9 Dynasty of Larsa Gungunum (1932-1906 B.c.), expeditions to Telmun brought back with other cargo gold, ivory, copper, carnelian, and silver.20 Babylon Sumu-ilum (1880-1845 B.c.), imports from Telmun included mesu wood, gold, objects inlaid with ivory, copper, and lapis lazuli.21 Dynasty of Larsa Rim-Sin (1822-1763 B.c.), silver, sesame oil, and garments were sent from Larsa to Telmun for copper. This text is from the middle period of RimSin's reign.22 Dynasty of Mari Yasmah-Adad(1796-1780 B.c.) writes to Hammurabi of Babylon about the delay of a caravan sent by the former to Telmun.23 Lexical series recovered in copies of the later Assyrian period but Mainly ijAR-ra: _hubullu
ibid., 180, 184, 220. 17 Leemans, Foreign trade, Magan copper, 19 (UET, mi, 1689); Magan onions, 21 (UET, mi,
16

751); Magan reed, 26 (UET, v, 678). 18 Leemans, Foreign trade, Meluhhla copper, 161 (UET, III, 368); A.AB.BA wood, 161 (UET, Im, 430, 660, 703, 752); mesu wood, 161 (UET, III, 818, 1498, 1241); ivory birds, 161 (UET, m, 757, 761, 764, 768, 770).
19 Leemans, Foreign 20 ibid., 24-6 (UET, 2t ibid., 27 (UET, v, 22 ibid., 36 (UET, v, 23 ibid., 138 (ARM, trade, 125, quoting OIP, XLIII, p. 194, no. 121. v, 526, 678). 292). 367). v, 14).

A PERIPLUS OFMAGAN ANDMELUHHA

557

which cuneiformists generally agree are derived from lists of the Larsa period or of the end of the Third Dynasty of Ur.24 wood of wood of Magan In, 155 A.AB.BA III, 204 MES Iv, 98 Magan chair Iv, 99 Meluhha chair InI,205 MES wood of Meluhha Meluha Iv, 194, Magan table III, 286 date-palm of Magan Magan copper xI, Iv, 195 Meluha table inI, 287 date-palm of Meluhha xx, Meluhha copper na4GUGc Meluhha (probably carnelian)
M1AGAN

We begin our inquiry with Magan and consider first the possible linguistic evidence for this location. Several toponyms from Arabia have been proposed as containing elements of the ancient name Magan. These include Ma'an and Maztin-'Uman,25although both present etymological difficulties which would bring the possibility of a direct development for either word into serious question. As noted above, Herzfeld and Eilers equate the region of Magan, in part, with Makran, a toponym designating the southern reaches of Baluchistan in both Iran and Pakistan. Now in Old Akkadian Magan may also read Makan with g and ki interchanging. The Babylonian version of the Bihistun trilingual inscriptions gives Ma-ak 26whereas the Old Persian is Maka.27 The people of Maka (ethnic Maciya),28 rendered Mykoi in Greek, are cited by Herodotus together with the Outioi (from Old Persian Yutiya) as forming a satrapy under Darius.29 This satrapy is the Maka of the inscriptions.30 The Yutiya have been identified by Marquart with the Jut nomads of south Kirman of the Islamic period.31 East of south Kirman is the geographical district of Makran in Iranian Baluchistan which is to be identified with Makan or Maka proper. Middle Persian texts give Makuran,32a form possibly influenced by the Baluch rendering which is the same. The spelling Makran is first attested in early Islamic texts. There appears to be substantial evidence, then, for a linguistic connexion between the toponyms Magan and Makran and this equation has now been generally accepted.33 There is no comparable justification to relate Magan with
For Magan and Meluhha chairs, 158; Magan and Meluhha 24 B. Landsberger, MSL, v. tables, 168; MES wood of Magan and Meluhha, 109; date-palm of Magan and MeluhBa, 117; A.AB.BA wood of MeluBBa, 105. For Magan and Meluhha copper see Matoul, LTBA, I, tablet 33, col. 5, 11.32-6. See Scheil, RA, xv, 1918, 115, for naaGUG Meluhha. 25 On these toponyms see Herzfeld, The Persian Empire, 63. 26 W. C. Benedict and E. von Voigtlander, JCS, x, 1, 1956, 3. 27 R. G. Kent, Old Persian grammar, New Haven, 1953, 117, 136, 141, 151.
28 ibid., 36, 'Maka is ethnic Maciya with the palatalization because the suffix began with the palatal sound '. 29 Herodotus, nI, xciii.

op. cit., 300. 91 See J. Marquart (ed.), A catalogue of the provincial capitals of Erinshahr, Rome, 1931, 77. 32 J. Marquart (ed.), Erandahr, Berlin, 1901, 16; A. Maricq, Syria, xxxv, 3-4, 1958, 336. " Herzfeld, op. cit., 63; Leemans, JESHO, nI, 1, 1960, 29; A. Salonen, Die Tiren des alten Mesopotamien, Helsinki, 1961, p. 99, n. 1.

30 See Herzfeld,

558

HANSMAN JOHN

either Ma'an or Mazun. The attraction for some writers who find Magan at least partly in Arabia apparently, has been to establish a kind of geographical compromise between the earlier and later texts. These, as we have seen, place Magan successively to the east and to the south of Mesopotamia. It would also not appear practicable that Magan should be situated on both sides of the Gulf of Oman during the period of its attestation as a region in the east. When Naram-Sin tells of subjugating Magan and of 'dispatching' Manui, Lord of Magan,34this certainly implies a single region with defined political limits rather than widely separated locations to be found in both Iran and Arabia. We examine next the early textual evidence for the situation of Magan. Naram-Sin and Gudea both claim to have quarried stone in the mountains of Magan for use in making statues. In the case of Gudea this stone has been identified as diorite.35 Manishtusu, the predecessor of Naram-Sin in Akkad, also brought stone from the mountains below the lower sea to be used for statues. The ships of Manishtusu reached these mountains by crossing the lower sea (the Persian Gulf) where the whole country as far as the silver mines was overthrown. It is usually taken that a crossing was made from the Persian side of the Gulf to parts of Arabia. However, Arabia is notoriously poor in deposits of silver or of lead ore in which silver is found.36 Moreover, Gadd has pointed out that the word designating 'crossed' in this text is of doubtful interpretation.37 The ships, in any event, could have crossed eastward from the mouth of the Euphrates to Makran. Here indeed, in the KEih-i Jabal Bariz north of the Jaz Miriyan depression above Makran are lead mines still worked partly for silver.38 A number of other lead mines located further north at the edge of the Dasht-i Lilt, also produce silver.39 It would appear more likely, therefore, that the silver mines indicated by Manishtusu were to be found in areas reached from Makran where they do exist and not from eastern Arabia where they do not. The same region whence the silver came, as we have seen, is indicated as the source of stone used for statues. This stone, like that brought from Magan for a similar purpose by Gudea, was possibly diorite. Now veins of metamorphic rock, including basalts, diorite, and steatite are found throughout the main mountain range, running east to west, in Iranian Makran.40 If our identification of diorite is correct, therefore, the Magan in Iran could have supplied both of these commodities to Mesopotamia in antiquity. What of the other imports from Magan ? According to our sources these include copper and MESwood. Deposits of copper ore have been recorded at mountains which rise just north of the Makran in KEih-iTaftan in the Sarh.add
p. 555. s6 The identification of na4ESI as diorite is deduced from the finding that nearly all the recovered statues of Gudea are of this stone. 36 R. J. Forbes, Metallurgy in antiquity, Leiden, 1950, 187. ** See

sTCAH, third ed., I, pt. n, 440.

38 R. J. Forbes, Studies in ancient technology,
40

vrI, Leiden, 1964, 212. *, J. Caldwell, Investigations at Tal-i Iblis, Springfield, Ill., 1967, 347-8.
CHI, I, 85.

A PERIPLUS

OF MdAGAN AND MELUHIUIA

559

Iran. Other deposits of copper are found at Saindak located north-east of KfRh-i Taftan. The Faryab district on the north-western side of Makranis also rich in copper ore.41 Dr. Ilya Gershevitch has demonstrated convincingly that MES mis wood of Or Sumerian and Old Babylonian texts is to be identified with the Old Persian yaki- in an Achaemenian trilingual inscription, and that the MESyakla-is in fact the sissoo tree.42 The sissoo is found along watercourses in both Iranian and Pakistan Makran.43 It is not, however, reported to grow in Arabia. Legrain suggests that Magan reed may be bamboo 44 although no reason is given for this identification. The Ur economic texts record that Magan reed was used for the long shafts of daggers and also as fittings for containers.45 Magan reed, then, was a plant of fairly strong structure. This does suggest bamboo and, indeed, in another of the Ur texts' six bamboos of Magan ' are listed among the stock of a carpenter's shop which producedfurniture and boxes.46 At the present time small patches of cultivated bamboo are to be found in the humid coastal areas of Makran in Iran. While it is not possible to determine with certainty the present botanic name of the Magan onion, one plant which may be indicated is discussed in Appendix I. In terms of other commodities associated with Magan and on the evidence of an acceptable linguistic connexion of the name, Makran in Iran satisfies all the requirements for this ancient country. Arabia, where some writers have placed Magan, does not provide comparable evidence.
MELUIIHHA

The toponym Meluhla attested in Sumerian and Old Akkadian texts was apparently intended by its Mesopotamian users to identify a country located beyond Magan. This is implied because in most texts where these toponyms occur in succession Magan is given first.47 As noted in the first section of this study, Landsbergerand Leemans suggest that Melubhais to be found to the east of the Persian Gulf, the latter giving Sind and the coast of western India. Later texts, of course, indicate a southern location but we are at present concerned only with the early occurrences of Melubha. As the older Magan and Meluhla are known to have exported copper and MES wood it is possible to consider that these countries with similar resources may have been neighbouring regions. In treating the location of Melugha, therefore, Baluchistan in Pakistan which lies immediately to the east of Yet does Makran/Baluchistan in Iran, perhaps merits our attention.
41
42

Caldwell, op. cit., 13, 375; see also CHII,I, 503.

I. Gershevitch, BSOAS, xIx, 2, 1957, 317-20. Calcutta, 1890, 13-14.

43 G. Watt, A dictionary of the economic products of India, III,

44 L. Legrain, UET, III, 859, 233. 45 ibid., 219, 493; 214, 363. 46 ibid., 260, 14. 47 On this point see Leemans, Foreign trade, 159.

560

JOHN HANSMAR

Baluchistan or some other region better suit as a possible source of the Meluhha trade with Mesopotamia? It will be useful at this point to review the early commercial links maintained between these two countries. We learn from an inscription of Gudea that ugu-wood and gold dust were Of 'brought down' from the mountains (KfJR) Meluhlhaand taken to Lagash. The deduced reference to mountains here 48 gives an important clue to the topography of Melulha. There are no mountains in the Indus plain which in its southern extent is Sind, sometimes identified with MeluBha. Eastern Baluchistan, on the other hand, is marked throughout its southern and central parts by trellised ridges of the Makran and Siahan ranges. These are replaced in the east by transverse ridges which run parallel to the western edge of the Indus plain. Baluchistan, therefore, would seem to fit this apparent and important topographical requirement for Meluhha. We do not know the exact identification of u~u-wood although ebony has been proposed.49 This tree grows in parts of India. Gold is found in India and was formerly reported in Afghanistan but not directly in eastern Baluchistan. Other commodities imported from Meluha were date-palms, MESwood, carnelians, A.AB.BAwood, copper, lapis lazuli, and multi-coloured birds of ivory. Of these, date-palms are found in southern Baluchistan where they flourish in the humid climate. We have seen that MES wood, that is to say the sissoo tree grows commonly in both eastern and western Baluchistan. Carnelians are found in the stream-beds of these regions. We turn next to the difficult question of GIS.A.AB.BA,a wood of Meluha. In Sumerian A.AB.BA means 'sea' and it is therefore possible to translate this word with its determinative as 'sea wood '. Now the armies of Alexander the Great encountered an interesting tree on their return to Persia from India, while advancing westward along the coast of eastern Gedrosia (the present eastern Baluchistan). Arrian describes these trees as having' a leaf like laurel, and they grow in places which are washed by the sea; the trees are left high and dry by the receding tide, but when the tide returns, they appear growing in the sea ... The trees here are sometimes even 30 cubits high and at that season they were in flower being very like a white violet, but of a very much sweeter perfume '.5o This is a fairly accurate description of the common mangrove (Ceriops candallearia) which grows in tidal shoals and stream outlets along the coast of Sind and on the eastern coast of eastern Baluchistan in Pakistan.51 Mangrove wood is
48 Although K R can be read as ' mountain ' or ' hill ', it may also indicate a land. In the case Gudea caused uiu wood and gold dust to be brought down (IM.TA.~) from there of KIR MELUJHHA, (Barton, The royal inscriptions, 184). In the case of KITRMAGANki, a more specific word for mountain, H1TR.SAG MAGANki, is also attested (Barton, op. cit., 190). (ed.), Une relation de la huitibme campagne de Sargon, 49 KB, In, 1, 37; F. Thureau-Dangin Paris, 1912, p. 53, n. 6. 50 Arrian, Anabasis, vI, xxii, 6-7. 51 Watt, A dictionary of the economic products of India, n, Calcutta, 1889, 261. For mangroves on the coast of Las Bela which geographically forms a part of eastern Magan, see M. A. Stein, GJ, cn, 5-6, 1943, 196.

A PERIPLUS

OF MAGAN AND MELUItJA

561

extremely hard, weighing 63 lb. per square foot. It is used in boat building, for house poles, and for making furniture at the present time.52 As we have seen, A.AB.BA wood was used to make chairs at Ur in the time of Ibbi-Sin and for a throne for Iluiuma of Assyria. Therefore if we accept that the association of with the sea may mean that this wood grows in sea-water and if GIS.A.AB.BA is to be placed in eastern Baluchistan, then the mangrove which grows Meluhha in salt-marshes and tidal creeks there, would seem to hold fair claim to be identified with A.AB.BA wood. There is no other tree in this region with a similar habitat. In regard to copper, deposits of copper-bearingore have been reported in the Ras Kuh and Pat Kuh areas of central Pakistani Baluchistan and slag heaps are to be found near Uthal Shah Bilawal which is located just east of the eastern Makran in Baluchistan toward Sind.53 The question of copper from Meluhha brings our discussion to an interesting and possibly valuable point. This is the percentage of nickel found in copper from the ancient Middle East. Analysis has shown that the nickel content of a number of ancient copper objects recovered in Anatolia, Palestine, and Egypt varied from 0 2 to 0419 per cent.54 Samples of copper from objects found in Mesopotamia which date from not later than about 2000 B.c., on the other hand, often show a nickel content of more than one per cent and occasionally as high as three per cent.55 Analysis of copper ore from Jabal Ma'adin in Oman (fig. 1) gives a nickel content of 0 19 per cent.56 The Jabal Ma'adin deposit is the only recorded source of copper in south-eastern Arabia. A low nickel content in this ore, therefore, would seem to argue against Oman as being a major copper supplier to Mesopotamia at least during the period under present consideration. Three lumps of crude copper recovered at the ancient Indus Valley site of Mohenjo-daro (fig. 1), on the other hand, were found to contain 0-31, 1-06, and 1 -27 per cent nickel.57 The copper used at Mohenjo-daromay have come from Baluchistan or possibly from Afghanistan where copper ore is also found. The source, in any event, would certainly have been one of these regions rather than Arabia. We may therefore suggest that the same sources were used to supply the copper to Mesopotamia, where samples dated to the earlier historical periods also show a high nickel content. This gives further support to our location for the countries of Magan and Meluhha, the main suppliers of copper for ancient Mesopotamia.
Watt, op. cit., II, 261. RGSP, I, 1, 1948, 10, and plate 1; vii, 2, 1955, 58. See also R. J. Forbes, Studies in ancient technology, vi, Leiden, 1964, 12. 54 C. H. Desch, ' Sumerian copper. Seventh interim report of committee appointed to report on the probable sources of the supply of copper used by the Sumerians' [Report, British Association for the Advancement of Science, annual meeting, Blackpool, 1936, Section H, 3 pp.; a copy is in the Library of the Institute of Archaeology, University of London]. 55 H. Peake, An., II, 8, 1928, 454. 56 ibid., 456. "7 Sir John Marshall (ed.), Mohenjo-daro, II, London, 1931, 484.
52 53

562

JOHN HANSMIAN

Multi-coloured birds of ivory are commodities from Meluhha listed in economic texts of the period of Ibbi-Sin. Yet elephants, which produce ivory, do not live in Baluchistan. However, a wisdom text dating to the Old Babylonian period indicates that neither were elephants to be found in Meluhlia. The text is given below. The donkey of Anshan The... of Parashe The cat of Meluhha The elephant of the steppe which . .. bites off the willow as if it were a leek '.A8 If Meluhha is eastern Baluchistan, as we have proposed, then the steppe should be identified with parts of India where elephants are found. Yet if this is so, why are ivory products associated with Meluhha in the economic texts ? One obvious answer would seem to be that Meluhha served as an entrep8t for the trading of commodities brought from regions which adjoined this country. Archaeological evidence tends to support this theory when it is applied to parts of eastern Baluchistan in Pakistan. There, at three different sites near the coast (Sutkagen-dor, Sotka-koh and Bala Kot on fig. 1), all of which may once have been accessible to sea-going ships, numerous pottery sherds of the Harappan culture of the Indus Valley have been recovered.59 Sites contemporary with these locations, but which are situated inland from them, on the other hand, show pottery of a related but separate culture.60 It has therefore been proposed that the 'Harappan' settlements on the Makran coast in eastern Baluchistan were outposts of the Indus civilization used as coastal stations or entrep~ts by Harappan merchants.61 The fact that ivory, but not the elephant which produces ivory is recorded as coming from Meluhha, supports this view if we identify eastern Baluchistan with Meluhha. That is to say that the Harappan merchants presumably would have brought ivory from India to sell to coastal traders in Meluhha. Such an arrangement would be advantageous to Mesopotamian seafaring merchants since an even longer voyage to Sind (fig. 1) or to the western coast of India would thereby be avoided. It is also to be noted, in this regard, that the estuary of the Indus is difficult of access for some months each year because of the annual monsoon winds and floods. The same need for convenient access to the sea trade would no doubt have brought gold from India and lapis lazuli from Afghanistan to the coast. Other items of commerce associated with Meluhha including date-palms, A.AB.BA wood identified as mangrove, carnelians, and copper, as we have seen, are found in eastern Baluchistan and no doubt would have been traded more directly at the coastal ports by the merchants of this region. Since all the commodities listed above were
G. Lambert (ed. and tr.), Babylonian wisdom literature, Oxford, 1960, 272. xxxvI, 142, 1962, 90-2; R. L. Raikes, AA, LxvI, 2, 1964, 290. 90-1. ibid., 92. W.

58

59 G. Dales, An., o60 Dales, art. cit., 61

A PERIPLUS OF MAGAN AND MELUIJHA

563

secured in Meluhia proposed as eastern Baluchistan, the early Mesopotamians would identify them collectively as being from that country. Indeed, as there is no one area in the region of our study where all the commodities associated with Meluhha are to be found in a natural state, the existence of an entrep6t appears to be an almost necessary deduction. Bearing in mind the need for Melulha to have mountains and to be located past the mountainous land of Magan, but before the country where elephants are found, it is difficult to identify a region other than eastern Baluchistan which would adequately suit these requirements. The possibility that western Baluchistan (now in Iran) and eastern Baluchistan (in Pakistan) were also divided politically during much of the period of our inquiry is supported by a study of the pottery recovered from sites in both regions. Most of the early pottery (periods I to VI of Bampfir (fig. 1), the main cultural settlement of antiquity in southern Iranian Baluchistan (western Makran) closely resembles that of Tepe Yahya (periods IV and V, an ancient site (fig. 1) in the adjoining Iranian region of south Kirman.62 The Bampir sequence also shows strong parallels with pottery of periods II to IV from Shahr-i Sfikhta, a site (fig. 1) in southern Sistan.63 As we have seen, on the other hand, the distinctive Harappan pottery found on three sites located near the coast in eastern Baluchistan identify these settlements as cultural outposts of the Indus Valley civilization. No objects associated with the Harappan culture, however, were recovered at Bamptir."4 Pottery of Bampfir periods V and VI does have some links with that of the separate Kulli culture of central eastern Baluchistan. This, in turn, shows Harappan influence. Yet the relevant pottery at Bampir is considered to be intrusive and not characteristic of the main sequence there."5 Pottery of the Kulli culture, while showing some affinities with that of the mature Harappan civilization in its latest phases, remains a separate sequence.66 Communication between inland eastern Baluchistan and the Harappan sites located near the coast there is evidenced by the presence of Kulli pottery at Sutkagen-dor, perhaps the largest of these settlements, and by the finding of some Harappan pottery at inland Kulli sites.67 At Shahi Tump, also in eastern Baluchistan but situated much nearer to Bampfir than are most of the Kulli sites, pottery of the two earlier phases indicate links with that of the Khiirab culture of Eastern Iran. However, the later phase at Shahi Tump shows strong links with the Kulli sequence.68 There is at present disagreement among excavators as to the dating of some of the pottery with which we are here concerned. On the evidence of similar sherds recovered at Tepe Yahy, the Bamplir sequence has been dated from the end of the fourth millennium to
62 C. C. Lamberg-Karlovsky,
63 64 65

Iran, x, 1972, 93.

M. Tosi, EW, NS, xx, 1-2, 1970, 13.

B. De Cardi, APAMNII, LI, 3, 1970, 268. ibid., 257. 66 Sir Mortimer Wheeler, The Indus civilization, third ed., Cambridge, 1968, 15-18. 67 On the Kulli sequence see Wheeler, op. cit. 68 W. A. Fairservis, Jr., The roots of ancient India, London, 1971, 227.

564

HANSMAN .TOHN

the middle of the third millennium B.c.69 On analogy with material excavated at Shahr-i Sfikhta, Bamptir I to IV is tentatively dated between 2700 and 2400 B.c. and Bamplir V to VI at between 2400 and 2000 B.C.70 The excavator of Bampfir suggests a terminal date of 1900 B.c. for the areas of that site so far examined.71 The period of the Harappan culture of the Indus Valley civilization is usually placed between 2300 and 1700 B.c.72 To sum up, and in spite of uncertainties of dating at Bampfir, the available evidence does suggest that the sites in western Baluchistan in these earlier periods remained culturally linked with settlements located to the east and north in Iran. The sites in eastern Baluchistan, on the other hand, show either direct or close affinities with those of the Indus civilization further east. As cultural links often reflect a political compatibility, although not necessarily political affiliation, our findings would seem to support the theory that western and eastern Baluchistan remained separate administrative entities, at least one from the other, during much of the period at present under study. This would allow Magan and Meluhha, as equated with western and eastern Baluchistan (Greater Makran)respectively, to retain their separate territorial identities in the earlier periods even though they may have been adjoining regions with similar natural resources. References to Meluhla as an eastern region disappear from contemporary Mesopotamiantexts at the beginning of the second millennium B.c. This implies a loss of direct contact which was probably encouraged, in part, by the rise of Tilmun (Bahrain) as a principal entrep6t for seafaring trade originating from regions located on or accessible to the Persian Gulf. In Sanskrit texts groups of lawless barbarianswho live beyond the Sindhu, i.e. to the west of the Indus, and who do not conform to Hindu customs, are identified as the Mleccha.73 The earliest occurrence of this name in Sanskrit is from the &atapathabrdhmana 600 B.c.) where the Mleccha are said to use (c. unintelligible speech.'74Now it has been noted by Gadd and Leemans that the oldertoponym Meluhhaand this wordMlecchashow certain similarities of form.'75 Moreover, the first reference given above would seem to indicate that these Mleccha occupied at least a part of the country where we have placed Meluhha. For the linguistic and geographical connexions of Mleccha see below, Bailey, pp. 584-6, ? A. By what name then was the country called Meluhha in the third millennium known during the later periods in Mesopotamian history ? To consider this question we turn to Assyrian texts of the early first millennium.
art. cit., 93. Lamberg-Karlovsky, communication from Dr. M. Tosi. 71 De Cardi, art. cit., 243. 72 Wheeler, op. cit., 132-3. 73 For the Sanskrit sources see below Bailey, pp. 584-6, (tr. J. Eggeling), SBE, xxvI, 32. 7' Satapathabrahman.a 76 Leemans, Foreign trade, 164.
70 Personal
69

? A.

A PERIPLUS OF MAGAN AND MELUIHHIA

565

Herzfeld has pointed out that a timber used by the Assyrian u.si-in-da-a, king Sennacherib (704-681 B.c.) for pillars and doors of his palaces,76is to be that is read' Sind wood'.77 In the same text Sennacheriblists i-mis-ma-kan-na, to say misu wood of Makan,78a timber also attested in other late Assyrian are and texts. Another product, an essence riQbaluhhu, its gum HILriQbaluhhi listed in Assyrian medical texts as medications used both externally and internally.'79Thompson has identified viQbaluhhuas galbanum, an extract of the plant Ferula galbaniflua which is native to Eastern Iran and Afghanistan.s80 This proposal is based partly on the finding that the different medical and occult uses of the modern galbanum are closely similar to those given for Tibaluhhu, and that both gum resins have the unusual property of being soluble in water. Now the word baluhhuwould seem an acceptable development of Meluhha with m and b interchanging according to an accepted linguistic association of these two consonants. A connexion of this kind, however, cannot be proven. Yet it may be useful for our discussion to note that on the basis of identifying with galbanum, this product would have been obtained by the <baluhhu Assyrians from regions located close to that of our proposed ancient Meluhha. Moreover,since overland trade from Eastern Iran and Afghanistan is not otherwise attested during the later Assyrian period, it is reasonable to suggest that commercial produce from those regions was shipped to Mesopotamia from entrep6ts situated on the Persian Gulf and along the northern coast of the Arabian Sea beyond. This agrees with the practice of importing F. galbaniflua to India from stations on this coast during the nineteenth century, Two points, as therefore, seem possible. First, "<baluhhu galbanum could have been exported from MeluBhha (eastern Baluchistan). Second, if Meluhha developed by the late Assyrian period to Baluhhu, this toponym may have become by association the name of a plant extract brought from there just as, for example, Magan reed and Magan onion were identified with the country of their procurement in an earlier period. A further indication of trade between Mesopotamia and eastern Makran and Sind in the late Assyrian period is perhaps suggested by references in Assyrian Or medical texts to an oil "A.AB.BA UKA.A.AB.BA, Assyrian kufabku. This substance, unidentified by Thompson, is mainly used externally as an astringent either in a poultice or applied as an oil directly on the body.s8 Now we have earlier proposed to identify Sumerian 6IS.A.AB.BA with the mangrove tree which grows along the coast of eastern Makran and Sind. The medical uses of
The annals of Sennacherib, Chicago, 1924, 106, 110. 77 Herzfeld, The Persian Empire, 70. 78 Luckenbill, op. cit., 106. 7, R. C. Thompson, A dictionary of Assyrian botany, London, 1949, 342-3.
80

7s D. D. Luckenbill,

ibid., loc. cit. For the habitat of this plant see Watt, A dictionary of the economicproducts
see For kcudablcu CAD, vn (K),

of India, In, 334. 81 Thompson, op. cit., 238-9 (IamA.AB.BA and ~amKA.A.AB.BA).

597.

566

JOHNHANSMAN

mangrove oil are closely similar to those given for uKA.A.AB.BA.82 It would

follow that the latter oil, presumably obtained from the tree of the same name, may be identified with a mangrove product. On the basis of our several proposals, therefore, we may suggest that the Assyrians of the period of Sennacherib had knowledge of Sind, Baluhhu, and Makran through the coastal trade with these regions. One may argue, on the other hand, that the Assyrian attestation of imisma-kan-na represented no more than the form of the Sumerian name of this wood which in Sennacherib'stime was actually cultivated in Assyria. According to this theory, Makan as associated with a wood may not necessarily have held a geographical meaning for the Assyrians. However, the Assyrians did use
UKA.A.AB.BA = ku~abku, proposed as mangrove oil, and the nearest habitat of

the mangrove from Mesopotamia is eastern Magan and Sind. Wood was also brought from Sind, a country not known from Sumerian or Old Babylonian texts. Therefore, since Sind lies east of Magan/Makran, it is difficult to believe that the Assyrians retained knowledge of the farther country and not of the nearer. The present discussion poses the question of why the toponym Sind is not attested in texts of the third and second millenniums B.c. One explanation, of course, would be that Sind is of Indo-Aryan derivation and therefore of late introduction into the country which it identifies. However, this cannot be proved. Alternatively it may be that while Meluhha, according to the writer's proposal, was primarily the country of eastern Baluchistan, the toponym may also have been applied at times to regions which extended to the east of there. The question remains problematical. What does appear to emerge by the period of Sennacherib,on the other hand, is a major renewal of the traditional coastal trade between Mesopotamia and the east; a commerce which included eastern woods, gum resins, and also, presumably, the lapus lazuli which Sennacherib used to decorate his palaces. Only two of these coastal countries in the east are known from Old Persian texts of the succeeding Achaemenian period. In inscriptions at Persepolis and Susa Darius I (522-486 B.c.) styles himself king of, among other regions, Maka (Makran)and Hidus (Sind) in that order.83 A politically unified greater Makran rendered Maka in Old Persian and extending to Hidus/Sind then, comprised a satrapy of the Achaemenian empire during the reign of Darius. It will be demonstrated in a later section of this study that the capital of Achaemenian Maka was located in the western Makran. Therefore, with government in the west and with the province unified, the Achaemenians would not need to make separate reference to Meluhha, proposed here as comprising in part the eastern Makran which is the southern part of eastern Baluchistan. The next satrapy east of Maka/Makran, as we have seen, is Hidus/Sind and the geographical proximity of these two regions exists to the present day.
82

83

Watt, A dictionary of the economicproducts of India, ni, 261. Kent, Old Persian grammar, 136, 141.

A PERIPLUS OF MAGAN AND MELUIIIUA

567

When the armies of Alexander the Great advanced westward from the Indus on returning to Iran, they passed through greater Makran. Alexander's historians identify this country as Gadrosia,s8 a toponym derived from the Gadrosii, a people whom the Greek army encountered upon entering the hill country of eastern Makran.85 Gadrosia is held by Herzfeld to be of Iranian derivation.86 Bailey suggests that Gadrosia is probably a Greek adaptation of hypothetical Old Iranian *gadravati-(see below, Bailey, pp. 586-7, g C). Very little is known of the political fortunes of Maka/Gadrosia during the Hellenistic period although Stephanus Byzantinus does record the name Alexandria Makarene.s7 This is an appellation probably given to Pura, the existing capital, in the west, of Achaemenian Maka where Alexander the Great rested for several weeks after his difficult march through Gadrosia. Arrian calls Pura the capital of Gadrosia.88 Makarene, on the other hand, is a properly formed eparchy name, not that of a satrapy as was Achaemenian Maka. Since the eparchy system is Seleucid, the toponym Alexandria Makarene,accordingly, is to be associated with the Seleucid era. Now by this later period Gadrosia or Gedrosia is still described by Strabo as bordering on Carmaniain the west and on the Indus in the east.89 This would suggest that the Maka of the Achaemenians and the Gadrosia of Alexander was divided into two administrative districts by the Seleucids. Indeed, Strabo refers to the tribe of Gedrosii at one point and to the Gedroseni as a people in a second reference.90 This latter term is the expected eparchal form of the tribal name. It is therefore possible that Strabo, in this instance, has adapted the Greek name of the governing eparchy of Gedrosia to identify the people who lived there. According to our findings then, the older Magan should be identified with Makarene in the west and Meluhha with our hypothetical 'Gedrosene' in the east. Now if, as suggested, Gadrosia is of Iranian derivation, this toponym would not have been associated with the country it identifies until after the Aryan invasion of Makran. Moreover, the Achaemenians, like their Greek successors, may have divided the satrapy of Maka into two administrative districts, Maka the centre of government in the west and in the east a district designated by an Old Iranian name from which the Greeks derived Gadrosia. At the period of Alexander the Great, therefore, the suggested most ancient toponym Meluhha and the related name of its people would have been replaced
84 Arrian, Anabasis, vr, xxii, 1. 8s5There are structural similarities of Gadrosii and Gadra, the name of a servile tribe who inhabit the Las Bela region west of Sind (D. Bray, The tribes of Baluchistan, Calcutta, 1915, 101). The language of these Gadri is 'local Sindhi vernacular' classed together with Jatki Sindhi (Census of India, 1931, Iv, 1, 143). It is not possible on the available evidence to consider a linguistic connexion of the two names. 86 Herzfeld, The Persian Empire, 335. 87 Stephanus Byzantinus, s.v. Alexandria Makarene. 88 Arrian, Anabasis, vi, xxiv, 1. 89 Strabo, xv, ii, 9 and 14. 90 Strabo, xv, ii, 9.

568

HANSMAN JOHN

for the passing Greeks by an Iranian designate of eastern Maka/Makran. On this hypothesis, it is evident that the Greeks would have identified the whole satrapy with the name of its eastern division (which they entered first) and the Achaemenians with that in the west. However, whereas the traditional (possibly Dravidian) Makan 91 (Maka) still survives in an adapted form (Makran), the presumably Iranian *Gadra- has not. If our several theories are correct OP *Gadra- would ultimately have been obscured as a name of the land and the majority of the residents of eastern Makran by a development of the far more ancient name Meluhha. The possible derivative of Meluhha will be discussed below. As a final note on Gadrosia we may even claim a statement of the late writer Marcianthat Patalene, the eparchy of Arrian's city of Pattala (in southern Sind) formed part of Gadrosia,92as evidence that in Seleucid or even Indo-Bactrian times the jurisdiction of Gadrosia was considered to extend as far eastward as the Indus. Almost nothing is known of the political-geographical situation in greater Makran during the Parthian period. Of the succeeding Sasanian era, al-Tabari records that Ardashir I (c. A.D. 222-40), the first Sasanian king, received the submission of the ruler of Makran.93 Sh5hpuhr I, the successor of Ardashir, lists Makuranand Sind as two of the countries under his rule.94 Both regions are also given in a late Sasanian geographical text as provinces of the Sasanian empire.95 This continued identification of eastern Makran with either Magan proper in the west or with Sind in the east during the four centuries' rule of the Sasanians would have maintained the political-geographical assimilation of the proposed region of Meluhha, which we have suggested was accomplished as early as the Achaemenian period. The Arab geographer al-Muqaddasi(A.D.985) gives two useful facts relating to the politically-united Makran of the early Islamic period. The first is that at that time Bannajbfir (the present Panj-gfilr),the capital of this province, was located in the eastern part. Al-Muqaddasi further states that this town was inhabited by Balisi (Baluchi).96 Here we have perhaps the first clear historical association of a people called Baluch with the proposed home country of Meluhha. Very little is known of the pre-Islamic history of these Baluch who speak an Iranian dialect and who now live, for the most part, in greater Baluchistan/ Makran. It has been suggested that the Baluchi settled in their present homeland with the name Baluch already associated with them. These proposals stem
91 See makcan, T. Burrow and M. B. Emeneau, A Dravidian etymological dictionary, Oxford, 1961, 304. of the outer sea (trans. W. Schoff), Philadelphia, 1927, 23-4; Arrian, 92 Marcian, Periplus Anabasis, vI, xvii, 2. 9* Al-Tabari, I, 820. 4 A. Maricq, Syria, xxxv, 3-4, 1958, 306, 336. 95 Marquart, fran~'ahr, 16. 96 Al-Muqaddasi, 478.

A PERIPLUS OF MAGAN AND MELUHHA

569

largely from an account given by Firdausi of the campaigns fought by the Sasanian king Noshirwin (Khosrau I, A.D. 531-79) against the Alans, the Baluch, and other rebellious tribes of the Caucasus.97 The original homeland of the Baluch, accordingly, has been located in the north-east of Iran. However, apparently unknown to supporters of this theory, Ndldeke pointed out in 1879 that Firdausi's narrative was not consistent with a similar account of this war given by the Arab historian al-Tabari who presumably had access to the same sources as Firdausi. While al-Tabari and Firdausi both list the Alans and several other tribes of the Caucasus,the Arab chroniclergives the Balangar for Firdausi's Baluch.9s Naldeke and Warner accept al-Tabari's text, as this is generally considered to be more reliable than the poetic narrative of Firdausi. Therefore, although the Baluch language does contain elements of north Iranian dialect, and although the people who use this language may have migrated to Baluchistan from more northerly regions, there does not appear to be evidence to suggest that they arrived in their present homeland identified as Baluch. A point put forward in support of the view that the Baluch came to Baluchistan and Makran after the period of the proposedly incorrect reference of Firdausi to Baluch in the north, is that a troublesome people, the Balflis,were found in south Sistan and in the Kirman desert by al-Muqaddasi during the latter part of the tenth century A.D.99These Balfis are equated with the Baluch and it has been further suggested that they did not arrive in the Makran until the Saljiiq invasion of Kirmin (A.D. 1041) obliged them to flee there.100 However, we have already seen that when al-Muqaddasi (before A.D. 985) visited Bannajbfir (Panj-gfilr),the chief town of Makran, the inhabitants there were Balfisi (Baluchi). Therefore the Baluch were apparently already the major inhabitants of the largest town in eastern Makran (our Meluhha and Baluhhu) at least 84 years before the Saljiiq conquest of Kirman. The Arabic Balfis is certainly Baluch with Arabic s used for Iranian ch, which is usual for this period. Yet the evidence suggests that their centre of residence by the second half of the tenth century was the Makran and not the Kirmnindesert where they were present in scattered bands. The name Baluch in this form is not attested in the pre-Islamic sources. It may be Iranian or non-Iranian (see below, Bailey, p. 586, g B). On the basis of our present discussion, Meluhha > Baluhhu > Baluch can be postulated, with m and b interchanging and with the Iranian ending ch replacing the nonIranian ending of Baluhhu. In this event, it is possible that the Iranian people now called Baluch could have adapted a form of the most ancient name of the country in which they settled, even though the name is now sometimes used in a non-ethnic sense. While the linguistic and historical evidence in support of a
97 Firdausi, Shhnsnma (tr. Warner), vII, London, 1915, 216-17. 98 T. NBldeke, Geschichte der Perser und Araber zur Zeit der Sasaniden,

Leyden,

1879

p. 157, n. 3.
** Al-Muqaddasi, 488-9. 39
100oo EI,

first ed., s.v. 'Bal165ist~in', 628.

PART 3.

VOL. xxxv.

570

JOHN HANSMAN

development of this kind is by no means conclusive, the theory remains plausible and attractive. Alternately, it may be suggested on different etymological grounds that Baluch derives from an Iranian origin. There are numerous possibilities (see below, Bailey, p. 586, ? B). Before drawing our conclusions, however, it is reasonable to review all the factors relating to the problem of Meluha. Whatever the case may be for Baluch, an Iranian derivation of this word would not invalidate the reasoning of the present article, based on different considerations, in regard to the location of Meluhha. After the Arab conquest of greater MIakran A.D. 710, much of this district in was administered as a unified province of the Eastern Caliphate. By A.D. 864 Ibn Khurdadhbih calls Makran a dependency of Sind. The Buwayhids, an Iranian dynasty, gained control of Makran after A.D. 945 and during much of the eleventh and twelfth centuries local rulers under the suzerainty of the Ghaznavids asserted considerable authority in the region. After this period of limited central control, greater Makran once again became attached in part or wholly at different times to different political forces from the north, the east, or the west. During the first half of the thirteenth century much of the Makran was dominated by the Khwarizm Sh~hs and later in the same century the Pathan sultans of Hindustan also occupied territory there. By the sixteenth century, greater Makranhad been divided between the Safawids and the Mughals. During the seventeenth century the Baluch and allied Brahui tribes shared political control in much of eastern Makran, and it is from this period of Baluch ascendancy in their own country that the term Baluchistan first comes (or returns in a developed form) into general notice as the name of a geographical region; 1o01 being, it will be remembered, after an this almost continuous foreign domination of the territory of Baluchistan for many centuries. Moreover, at this same period in Safawid Iran, the name Baluchistan is first associated in surviving records with western Makran where the Baluch also increased their political hold.'02 The Baluch continued to war with rival tribes and against other invading forces in these territories until the middle of the eighteenth century when almost the whole of the Makran and the more northern parts of Baluchistan in present-day Pakistan were incorporated into the Baluch-dominated Khanate of Kalat. Western Makran was gradually absorbed by Iran after the rise of the Qajar dynasty. The present frontier between eastern and western Makran was delimited in 1870-2. Accordingly, the proposed ancient and traditional separation of the two regions has been re-established. With the final division of this territory the name here considered to have been traditional also continued in use, but now possibly because of long periods of political union both Makranand Baluchistan are associated with both regions.
1o0 For a concise account of the Islamic history of Baluchistan and a list of the relevant sources, see ibid., 634-9. o10 B. Spooner, 'Notes on the toponymy of the Persian Makran', in C. E. Bosworth (ed.), Iran and Islam, Edinburgh, [1971], 520.

Ci72

Ci72

~i~iiiiii!

iii ~;'::li:~p?.::::LA

::~l::: : ?:: ii~i~-ii~iliiiii ex i~"ii xWi ?i- E:i::::::::i-_:::ii'::''ii:ii~:::: i?i:~ ~. i,W ::-; ::ii:::~ i .il ~~Kill il~lll~-iii:~~i::lii ,:: :::ii"~ iiiii: ~ OW N.::--?:::::
j
l

:- il,

iilk

,~-,;-i_::

''W o,:-'

: ::':::' ::::''-::: :::-: i::::::::::::_::::: ol:-,_,:i -:::i: ~1?fill;-:

-i:1'8' ' ~~ii il~l
?i ot~ ':i

VIEW OF BAMPfR

FORT FROM THE SOUTH-EAST

A PERIPLUS MAGAN MELUIHLA OF AND

THE CITY OF MAGAN

571

The question now arises of the location of the main administrative centre of Magan (western Makran). This would have been the residence of Manui, Lord of Magan, whom Naram-Sin (2291-2255 B.c.) conquered. It may, however, be helpful to our inquiry to consider first the residence at a later date. According to Arrian, the army of Alexander the Great advanced westward through Gedrosiawith much suffering for lack of food and water. After 40 days spent partly in waiting for supplies, Alexander arrived at the district of Pura which was the capital of Gedrosia.103 Since Gedrosia is identified with the Achaemenian satrapy of Maka, Pura would have been the capital of this province. From Pura Alexander moved directly into the satrapy of Carmania (Kirman).'04 Thus the chief town of Maka/Gedrosia is to be located in western Maka, where we have placed Magan. This Pura has been associated by Holdich and Tarn with the town of Fahraj (now Iranshahr) which is located in the Bampflr valley above the eastern end of the Jaz Mflriyadepression.105 However, Stein found no substantial archaeological remains at Iranshahr. A further identification of Pura which the writer favours is Bamphir(fig. 1), a town and cultivated area of the Bampairvalley located 25 km. west of Iranshahr. The toponym Pura is perhaps contained in the second syllable of Bampiir with the final a lost in development to Middle Persian. Bam- is possibly an addition or may always have been present but not emphasized by the Greeks who were more concerned with the district name which may only have been Pura. We may indeed have Bam (of the district of) Pair(a)to distinguish this settlement from another chief town, Barn, located in a Kirman district also called Barn, which is situated 300 km. to the north-west of here. Now the valley at Bampaircontains the most extensively irrigated agricultural tract of the whole of central, western Makran. This area has apparently been prosperous for many thousands of years, for a large archaeological site measuring about 90 metres across was discovered there by Stein.'o6 Both Stein and De Cardi have carried out excavations at Bampair and De Cardi has established a sequence of six successive phases for the pottery in those parts of the site so far examined. As we have already seen, according to the evidence of similar pottery recovered at Tepe Yahyy and at Shahr-i Sfikhta, the recorded Bampair sequence has been dated either from the early fourth to the middle third millennium B.C. or from about 2700 to 2000 B.c. Whichever of these estimates may prove more accurate, there is clear evidence of a substantial cultural occupation at Bampfir in the third millennium. This brings us to the matter of the great mud fort of Bampir located 200 metres to the east of the
103 Arrian, Anabasis, vI, xxiv, 104 See Appendix II for a note 105 T. H. Holdich, GJ, vii, 4, 1948, map. 106 M. A. Stein, Archaeological

1; vi, xxvii, 1. on Alexander in Carmania. 1896, 409; W. W. Tarn, Alexander reconnaissances, London, 1937, 105-6.

the Great, II, Cambridge,

572

JOHN HANSMAM

excavations there (fig. 2). The fort sits on a mound which measures approximately 60 metres long by an average 30 metres wide (plate I). A lower section of the fort defences and mound rises 15 metres above the valley floor. A higher part of the mound, on which the main structure of the fort is built, extends a further 14 metres above this level. Stein noticed that natural gravel deposits are exposed at the sides of the lower part of the mound. Yet no such deposits were found nor was any pottery seen by him on the slopes of the upper area. Because of this Stein suggests that successive collapses and rebuildings of the mud fort through the years may now deeply overlay any early cultural remains which may exist there.'07 Indeed, as Bampfir is the largest agriculturally productive area in the central Makran,and as it is quite possibly to be associated with Pura, then we should expect that remains of a number of successive occupations could be sealed under the great mound of the fort. There is no clear geological reason why the upper mound should stand up in sharp isolation from the partly natural lower mound (plate I and fig. 2). There are no other high mounds of this kind in the vicinity. Only excavation can finally answer this question, of course. Yet the evidence for locating Maka/Pura here and also, therefore, Alexandria Makarene, is sufficiently encouraging to make further archaeologicalinvestigation of this enormousfeature a most intriguing prospect. The previous excavations at Bamplir have confirmed that the thirdmillennium occupation there is far more extensive than that of any other site of this period presently known from western Makran. Moreover, Bamptir is located on the major east-west track which passes from Sind through Makran and to Kirman province beyond. It is also situated on the most direct route leading from the region of the great site of Shahr-i Siikhta in southern SIstan to the large natural bay of Chah Bahar (fig. 1) which lies almost directly south of Bamptir on the Gulf of Oman. It would seem, then, that ancient Bampfir may hold a reasonable claim to have been the site of the residence of Manui, Lord of Magan, and therefore also the location of the chief town of the country of Magan identified here as western Makran. During the early Islamic period, the main port of greater Makran was Tiz, a known site (fig. 1) located on the eastern side of the bay of Chah Bahar.ls08 Al-Muqaddasi(A.D.987) mentions the great warehousesof TIz and its population of all nations.'09 Al-Idrisi (A.D.1154) records that Tiz was much frequented by vessels from Fars and Oman.110 Tiz was probably also the main port for the commerce of Sistin and for the lower Helmand valley of Afghanistan which are located directly north of the bay of Chah Bahar. If this superb natural harbour held an equal importance for the sea trade in antiquity, we may expect to find somewhere along its length, perhaps under Islamic Tiz, the remains of
110. the ruins of Tiz, see Stein, op. cit., 87-8. o109 Al-Muqaddasi, 478. 110o Al-Idrisi, Nuzhat al-mushtiq, translated into English by S. M. Ahmad as India and the neighbouring territories, Leiden, 1960, 46.
o107 ibid., 108osOn

A PERIPLUS OF MAGAN AND MELUIJIA

573

.AC

Fo

Scale 0 100 200 300 metres

3 Contoursat approximately metres (after Stein) excavations A 'Archaeological

FIG. 2. IBalnlir

574

JOHN HANSMAN

a major port of Magan. It would have been to this port that the ships of Ibbi-Sin and those of Gudea and of other rulers in ancient Mesopotamia once came for stone, copper, and wood."l
THE SOUTHERN MAGAN AND MELUHHIJIJA

The most recent text in our list giving evidence of direct contact between Mesopotamia and Meluhha in the east is ascribed to Gudea of Lagash (c. 2220 B.c.). Over 200 years later, during the reign of Ibbi-Sin (2029-2006 B.C.), officials at Ur arranged for the exchange of garments and wool for copper from Magan. By the middle of the reign of Rim-Sin (1822-1763 B.c.) of Larsa, surviving texts indicate that exports from Mesopotamia to the east were usually shipped to Telmun (Bahrain)and there exchanged for other commodities. These included copper, mesu-wood, carnelian, and ivory, all of which had been supplied earlier by Magan and Meluiha. We may therefore adduce that Telmun (fig. 1) now served as an entrep6t for articles of commerce which were brought there from the traditional sources. The Telmun trade in the west continued at least until the time of Yasmah-Adad (1796-1780 B.c.) of Mari who speaks of a caravan sent from his city to Telmun, in a letter to Hammurabi of Babylon. After this period and that of the reign of Yasmah-Adad's part-contemporary Rim-Sin, mentioned above, references to Telmun do not again occur in the recovered Akkadian texts for nearly 400 years. The available evidence suggests a suspension of the eastern trade with Mesopotamia at some point during the eighteenth century B.c. Whether or not a disruption did occur and if so, whether it can be at least partly ascribed to events which destroyed the Indus civilization at about this same period, is a question which cannot be answered in the light of present knowledge.112 The name Meluha or Miluha reappears in letters written by Rib-Addi, regent of Egypt at Gubla (Byblos), to Amenophis III (1411-1375 B.c.) and to Amenophis IV (1375-1358 B.c.). In a series of urgent messages, Rib-Addi pleads for troops from Misri (Egypt) and from Meluha to be dispatched to Gubla for the protection of Egyptian territorial conquests on the Eastern Mediterranean."3 In other letters Rib-Addi asks the Egyptians to send men of Kashi to assist in this endeavour."' Sama-Adda, an official in south Palestine and presumably a near-contemporary of Rib-Addi, also asks the Egyptian king for men from Kashi."5 At about the same time Abdi-Hiba, regent in Jerusalem, complains to the Egyptian king that the Kashi people stationed at Jerusalem had committed offences against him."6
111 See p. 556.
112 On the end of the Indus civilization 113aS. A. Mercer (ed.), Tell el-Amarna tablets, Toronto, 114 ibid., I, 422, 436.
115

see Wheeler, The Indus civilization, 132-3. 1939, I, 266, 366, 374, 382, 390, 440.

ibid., I, 222.

16 ibid., I, 712.

A PERIPLUS OF MAGAN AND MELUIHHA

575

Now the toponyms Kas and Kash occur in Egyptian inscriptions of Sesostri I and it has been demonstrated by means of parallel inscriptions that this is in fact Kush."17 Under the Middle Kingdom the political district of Kush to the Egyptians comprised the old Nubian province of Wawat; this was located between the first and second cataracts, and it also included the more southerly region of Kush proper down to the area of the fourth cataract in the Sudan."8 In Neo-Assyrian texts, to be discussed below, an identification of Kush with (the southern) Meluhha is indicated. Rib-Addi always uses Misri, a form of the indigenous name for the territories of homeland Egypt, to identify that territory. However, in many letters he appears to have resorted to a substitute name, Meluha, for that of the indigenous toponym Kashi or Kush which he and officials in neighbouring Palestine also knew and sometimes used in place of Meluha. Before commenting further on these findings it will be useful to complete a chronological survey of the later occurrences of Meluhhia,Magan, and Telmun. A letter from an official of Kassite Nippur of the time of Burnaburiash II (1375-1347 B.C.), who was stationed on the island of Telmun (Bahrain), shows Telmun to have supplied dates to Mesopotamia at that period."9 No other items of commercefrom the once flourishingentrep6t of Telmun are listed in the recovered Kassite texts. This suggests that the trade between Telmun and the eastern countries of Magan and Meluhha, which seems to have lapsed already by the eighteenth century B.C.,remained largely suspended. On the other hand Burnaburiash, a contemporary of Rib-Addi, and other Kassite kings are known from the Tell el-Amarnatablets to have sent gifts of lapis lazuli to Amenophis IV in Egypt.120 This implies that lapis, which in Asia occurs only in Afghanistan, was imported into Mesopotamia at that time. We do not know whether this commodity was obtained from coastal trading with the eastern ports or brought overland direct from Afghanistan, as had been the case in the time of Enmerkar of Uruk (c. 2700 B.C.). After the period of Rib-Addi, an association of Meluhha with Nubia and the Sudan is next recorded in an Akkadian text from the Hittite royal archives at Boghazkoi. This relates that slaves of Meluhibawere among gifts received from Egypt by an unnamed Hittite king.'21 We hear next of Meluhha and Magan from inscriptions of the Assyrian king Tukulti-Ninurta I (1244-1208 B.C.). In one text he styles himself King of Assyria, King of Karduniash, King of Sumer and Akkad, King of Sippar and Babylon, King of Telmun and Meluhl4a,King of the upper and the lower seas.122 The detailed Assyrian accounts of the conquests of Tukulti-Ninurta make
G. Posener, Kush, vI, 1958, 45-7. For discussion and sources on Kush see CAH, third ed., II, pt. I, 346-9. 11"9 JCS, vi, 4, 1952, 143 f. 120o Mercer, Tell el-Amarna tablets, I, 22, 70, 80-8, 128-44. 121 KUB, III, no. 51, 1. 20, and no. 52, 1. 6 f. 122 E. F. Weidner, A f. O, Beiheft 12, 1959, 30; ARA, I, 59.
117 118s

576

HANSMAN JOHN

no mention of his invading either the eastern or the southern Meluhha nor do they indicate that he at any time reached Telmun. This suggests that 'King of Telmun and Meluhha ' and 'King of the upper and the lower seas' are to be regarded as formula titles, that is to say as traditional, honorific designations which had no real political meaning in Mesopotamia. In the case of'King of Sippar and Babylon', the title would represent an archaic retention, for in the same passage Tukulti-Ninurta uses 'King of Karduniash', the contemporary Assyrian name for Babylonia. In a second text Tukulti-Ninurta boasts of overthrowing the land of Shubari and the Nairi lands up to the border of Magan.123 The first two regions given in this account are located north of Assyria. The placing of Magan in the vicinity of these territories suggests that the author of the text confused the location of that country. Until this date only Magan in the east has been indicated in the surviving sources, the toponym being last attested in economic texts of the Third Dynasty of Ur dated to the reign of Ibbi-Sin (2029-2006 B.C.). According to one theory, an eventual awareness of this inaccuracy by Assyrian officials resulted in Magan being dropped from Tukulti-Ninurta's later accounts of his conquest of the Shubari and Nairi
lands.124

MeluBha does not occur again in the historical texts for 500 years. It reappears in the annals of Sargon II (721-705 B.C.) where, during an Assyrian advance on the rebellious state of Ashdod (in Palestine), the king of Ashdod fled to the side of Musur (Egypt), which is on the border of Meluhha.125 For the first time now there appears to be clear evidence of the association by the Assyrians of Meluha with a district in the south near Egypt. An inscription of Sennacherib (704-681 B.c.), successor of Sargon II, relates how rebels who rose against Assyrian rule in Palestine sought military assistance from the king of Musur (Egypt), and bowmen, chariots, and horses from the king of Meluhha. Sennacherib claims that he later captured the charioteers of Egypt and Meluhha.1'26 The succeeding Assyrian king, Esarhaddon (680-669 B.c.), claims in one text to have conquered the countries of Tyre, Musur, and Kush in that order.'27 According to another text, Esarhaddon advanced from Tyre to Musur and then toward Meluhha.128 This implies that Meluhha is to be found beyond Musur (Egypt), that is to say the country of Kush attested in the first reference. In a third text, Esarhaddon styles himself' King of the Kings of Musurand Kush' and in a later passage as ' King of Musur who bound the King of MeluhBa'.129 In the last instance Kush does not occur among the titles of Esarhaddon. It
123 Weidner,
124 125

21. CAH, n, third ed., pt. n, fasc. 49, 14.

126 Luckenbill,
127 128 129

E. Winckler, Die KeilschrifttexteSargons, I, Leipzig, 1889, 114, 116; ARA, ni, 32. The annals of Sennacherib, 31-2.

iR. Borger, Die Inschriften Asarhaddons, Graz, 1956, 86.

ibid., 112, ibid., 36.

A PERIPLUS

OF MAGANAND MELUHIIA

577

seems clear from the above comparisons that Kush and Meluhha are used alternatively to identify the same country. In a fourth text, Esarhaddon resorts to using a number of traditional titles for himself. These include the ancient appellation' King of Sumer and Akkad ' and also 'King of the Kings of Telmun, Magan, and Meluhha '.130 This account deals only with the early reign of Esarhaddon-he gives specific reference to his first year-and was apparently inscribed before his invasion of Egypt. Moreover, since there is no evidence that Esarhaddon at any period during his reign invaded Telmun or the eastern Magan and Meluhha, the use of a title which includes these toponyms suggests an honorific association which carries no real political meaning. This possibly reflects a convention which we have earlier ascribed to Tukulti-Ninurta and may have originated with Sargon of Akkad's claim to have caused the boats of Telmun, Meluhha, and Magan to be moored at the wharf before Agade. Lastly to be considered in the present survey are the annals of Assurbanipal (668-627 B.c.). In editions B and D of this text Assurbanipal claims that 'on my first campaign I marched against Magan and Meluhha '.131 Now the very next line states that Tarqfl, King of Musur(Egypt) and Kusi (Kush) had revolted against the Assyrians. Troops of Assurbanipal reoccupied Egypt. Tarqfl meanwhile had fled to Kusi, but soon again became a menace to the Assyrians.'32 The second campaign of Assurbanipal, accordingly, was also directed against Musur and Kusi.'33 In the first line of the first text treated above, it is to be seen that Magan and Meluhhaare given in a formalized introductory statement. Yet the passages which follow refer consistently to Musur and to Kusi. It is therefore reasonable to consider that Magan is used here as an alternative toponym for Musur and Meluhha for Kusi. We have determined that the association of Meluhha with the name Kusi or Kush dates at least from the period of Rib-Addi, the early fourteenth century B.C. Over 600 years later, this identification for Meluhha was used by the later kings of Assyria. However, only Assurbanipal appears to have equated Magan with Musur(Egypt). What factors brought about the association of these toponyms, originally applied to regions lying to the east of Mesopotamia, with other lands in the south-west ? Several possibilities have been advanced. We have, of course, the theory proposed by Landsberger whereby a namechange occurred because gold and ivory once supplied to Mesopotamia by the eastern Meluhhla,were later to be secured from Ethiopia through Egypt. It is to be noted, in this regard, that Landsberger used the classical appellation of 'Ethiopia' which would have included the country of the Sudan. Yet a parallel
ibid., 80. A. Piepkorn, Historical prism inscriptions of Ashurbanipal, text D.31, no. 12; ARA, II, 292. 132 Piepkorn, 30-8; ARA, II, 292-3. a13Piepkorn, 38; ARA, xI, 295.
131 130

Chicago, 1933, text B.30 and

578

HANSMAN JOHN

in traded commodities from the east and the south-west would hardly seem to justify, as Kramer points out, a purposeful name-change of this kind for such important regions as these.134 Alternatively, it would seem difficult to propose two indigenous Meluhhas; for while there is some possibility of a development of Melutha > Baluch in the east, in the south-west Egyptian texts do not attest this toponym. It is almost certainly derived in Mesopotamia from the Akkadian usage of a possibly non-Akkadian word. Is there a third possible explanation for the sudden appearance of a Meluhha in the south ? It will perhaps be useful, in this regard, to consider the evidence of an Akkadian geographical treatise, the existing copy of which is thought to have been inscribed at about the period of Sargon II of Assyria.'35 This tablet lists numerous regions associated with the conquests and period of the much earlier Sargon of Akkad and presumably is based ultimately on an original text prepared during his lifetime. Line 30, which holds immediate interest to our study, states: '" 2 uv bereli-id-du iitu KUN narPu-rat-tiadi pit matMe-luh-hu Md-ri i. In translation this reads,' 120 berufrom the water-basin of the Euphrates to the border of Meluhha [is] Mari '. The sentence seems to indicate that the border between Meluhha and Mari is to be found 120 bgru (1,283 km.) from a point near the Euphrates. The water-basin here apparently is the ancient complement of the large, shallow lake-the present Haur alexists at the lower extremity of the modern river (fig. 1). H.ammar-which known to have emptied into a shallow lake or swamp at least The Euphrates is from the period of Sennacherib (701-681 B.c.),137and there is no reason to suppose that this was not also the case during the reign of the first Sargon. We turn next to the land Md-ri which our text places between the Euphrates and Meluhha. If within the early context of the geographicaltreatise, we accept that the most ancient Meluhhain the east is indicated here, the related district of Md-ri cannot be the Mari located on the Euphrates in north-western Mesopotamia.138 Two Akkadian texts from Nippur which probably date from the early second millennium B.C., may prove more helpful to our inquiry. These treat of the conquests of the first Sargon and both mention Ugu, the man of Mari and Elam, who stood before Sargon, king of the countries of the earth.'"39 As this Mari is associated politically with Elam, it can only be located to the east of Mesopotamia in Iran where Elam is to be found. Accordingly, as Meluiha of the period of the first Sargon was also located in the east, we are justified in measuring a distance of 1,283 km. (120 beru)140east from the eastern limit of the Haur al-Hammar. By following the traditional overland route
134 See above, 135 Weidner,
13s

p. 555, n. 11. A f. O, xvI, 1, 1952, 1. The annals of Sennacherib, 74; ARA, I, 146.

ibid., 19.

'37 Luckenbill,
a138 139 140o

See ARM, xv, 1954, 129, for text references to this Mari.
Barton, The royal inscriptions, 105-6 and 112. The b ru has been reckoned as equalling about 10 69 kilometres. See RA, xviII, 1921, 133.

A PERIPL

US OF MIAGAN AND MELUHJA

579

through Khuizistan, Fars, and south-eastern Iran, we arrive at a point which is very close to the present border between Iranian Makran and Pakistani Baluchistan (fig. 1). This, then, would be the region of the border of the eastern Mari (which may represent Magan) with Meluhha. It is the same region where we have postulated the frontier between the eastern Magan and Meluhha. The present evidence, therefore, would appear to add a valuable confirmation to our theory regarding the situation of earlier Meluhha. In developing our inquiry, we turn again to the letters of Rib-Addi which give the earliest record of an association of Meluhla with Nubia and the Sudan. Why, one may ask, did this official use the name Meluhha when in other letters he and near-contemporariesof his in Palestine give 'Kashi', a form of Kush, the indigenous name of much of the region indicated ? To begin with, we should remember that Meluhha, as associated with the south-west, occurs only in Akkadian texts. This brings us to consider the possibility that the writers of these documents in provincial Gubla (Byblos) were following a convention which presumably would already have been adapted by the more knowledgeable Akkadian scribes of Assyrian and Kassite Mesopotamia. It would seem equally acceptable that scribes living in the scholarly centres of Mesopotamia at the time of Rib-Addi, had access to copies of the geographical treatise, mentioned above, which is a transcription of a much later date. This being evident, we may legitimately ask what interpretation would have been given in these centres and in the light of contemporary geographical knowledge, to the passage of our text concerned with Mari and Meluhha. Firstly, it is entirely possible that scribes in Mesopotamia during the middle of the second millennium B.C.would not have been aware of the Mariin southern Iran. The association of this Mari with Elam, as we have seen, is at present known only from two texts found at Nippur which probably date from the very early second millennium B.c. Moreover, if the eastern Mari was indeed unattested at this later period, it is conceivable that Md-ri of the geographical treatise may not have been identified with any specific region. In this event our referenceto the land of Maricould have been subject to the broadest possible interpretation. We turn now to Meluha. It will be remembered that, in this form, the toponym is last associated with a country of the east in an Assyrian economic text dated to the reign of Iluiuma. This is about 500 years before the period of Rib-Addi. For 200 years after the reign of Ilubuma, economic texts from south Mesopotamia attest continued commercial ties with Telmun, which then presumably served as entrep6t where traded goods once brought from Magan and Meluhha were obtained. A direct loss of contact between Mesopotamia and these farther countries by this later period is suggested. Not until the reign of Sargon II (721-705 B.c.) and that of his successor Sennacherib (705-681 B.c.), do Assyrian building inscriptions and medical texts indicate, through the associated names of certain plants and woods, that commerce had again been established between Mesopotamia and the eastern

580

HANSMAN JOHN

lands Magan and Meluhha (proposed Baluhhu). In view of the absence of in contemporary references to a Meluhhla the east for over 1,200 years, it would seem reasonable to suggest that the exact location of this country may gradually have become lost to the Mesopotamians. At the same time, it is known from a number of the el-Amarna tablets that gold and ivory, once obtained from the eastern Meluiha, by the middle of the second millennium B.C.were sent in quantity by Amenophis III to KadashmanEnlil, Kassite king at Babylon, and by Amenophis IV to Burnaburiash II (1375-1347 B.c.) a later Kassite ruler there.'41 In one letter, Burnaburiash asks the Egyptian king to send him trees and plants made of ivory and coloured.'42 This is strikingly reminiscent of an economic text of the period of Ibbi-Sin in which multicoloured ivory birds of Meluhhlaare listed. In any event, the gold exported by the Egyptians may have come from Nubia (Kush) where it is found and the ivory from lands located further to the south of there but acquired by way of Kush, the southern Meluhha. Now Landsberger, as we have seen, suggests that the Mesopotamians of the second millennium B.C.may have learned, presumably from a lexical series or from archival texts, that in earlier periods gold and ivory were brought from a country called Meluhha to Mesopotamia. There is no direct evidence to support this view; nevertheless it remains hypothetically plausible. Therefore, if this were the case and if the location of the most ancient eastern Meluhha was indeed no longer clearly known in Mesopotamia at the period of our present inquiry, the possibility of a location in the south-west by association with goods brought from Kush could have arisen. That the late Assyrian kings were not content with using Mielulhaexclusively as the conventional identification for Kush is indicated by their frequent substitution of a form of the indigenous name. By the reign of Assurbanipal, this usage became entirely formalized in that it is restricted to introductory statements given to identify the regions of Assyrian campaigns. Yet for the actual descriptions of those campaigns, Kush continued to be used. Therefore, when Assurbanipal substituted Meluhha for Kush he presumably intended this to reflect a traditional Akkadian usage. Possibly, however, he and his predecessors in Assyria confused the region of association. Given the evidence which may have been available to the scribes of this period, it was perhaps a reasonable assumption to make. It is doubtful that the later Assyrians would have continued to accept an identification of Meluhha with Kush in the south-west if, through renewed trade, they regained direct knowledge of the older Melulhhalocated to the east of Mesopotamia. This is a point to which we will give some attention. It has been noted that from at least the period of Sargon II and Sennacherib the Assyrians were aware of the countries of Magan and Sind in the east.
14' Mercer, Tell el-Amarna tablets. For gold see I, 10, 16, 24, 32, 44-6; for ivory, I, 16, 41, 54-6.
142

ibid., I, 36.

A PERIPLUS

OF MAGAN AND MELUHIIA

581

Moreover, we have suggested that the Assyrians may also have had knowledge of the territory of eastern Makran (MeluIhha), proposed to have been identified by that time with the plant-associated name baluhhu. A development Meluhha > ' Balubhu 'is linguistically acceptable. Accordingly, if we postulate a Baluhhu in the east, the Assyrians would not necessarily associate this toponym with a MeluBhawhich at some earlier date they had applied to a region in the south-west. Furthermore, if Baluhhu is the developed form, then this would confirm that the Assyrians had borrowed the archaic rendering from the older Sumerian and Akkadian texts, to identify Meluhha in the south. Therefore, as these most ancient texts are concerned only with Meluhha in the east, the equation with Kush can be shown on the above proposal to be based on misunderstood sources. Also, when the king of Meluhha (here Kush) returns the fugitive king of Ashdod (in Palestine) to the Assyrians, Sargon II relates that the fathers of the king of Meluhha had not ' since the far-off days of the moon god's time' sent messages to the 'kings, my fathers '.143 This passage gives ample evidence of a long break in contact between the Assyrians and a land known to them, presumably from earlier sources, as Meluhha. The reference also supports the theory that Baluthu to the later Assyrians, whom we propose to have maintained trade contacts there, was no longer identified by them with the toponym Melutha. Therefore, the Assyrian equation of Meluhha with Kush, based partly perhaps on a comparison of the geographical treatise with the available lexical series, may have resulted from a logical but incorrect interpretation of these earlier records. The later Assyrians, as we have seen, most probably did retain direct knowledge of Magan in the east through trade from there or at least from Sind located farther to the east. In Assyrian annals only one text, in three surviving copies, appears to equate Magan with Musur (Egypt). This is the inscription of Assurbanipal, already noted, which gives Magan and Meluhla in an introductory line of a passage otherwise wholly concernedwith the lands of Musurand Kush. The collected evidence suggests that the single attestation of a southern Magan was brought about by Assurbanipal's adaption of an Old Akkadian formula title which included these toponyms. Indeed, it was perhaps more essentially the identification of the Meluiha half of this designation with Kush which resulted in what may possibly best be called a circumstantial, late Assyrian association, of the other half with Egypt.
APPENDIX

' Magan onion' Although any identification for 'Magan onion' must be based on hypothetical reasoning, the question is of interest for the location of Magan and merits some consideration.
143

Winckler, Die Keilschrifttexte Sargons, I, 116;

ARA, n, 32.

582 Sumerian SUM.SAR

MA.GANki,

JOHN HANSMAN

attested in a text of the period of Ibbi-Sin

(2029-2006 B.c.), can mean 'onion' or 'garlic'. Yet it would seem unusual that either of these plants, both of which are cultivable in Mesopotamia, should be imported at great cost from far-off Magan and that produce of this kind should be delivered to the Nanna temple at Ur, probably as a tithe.144 On the other hand, there is a plant, Ferula alliacea (one of the giant fennels), which is found in southern and eastern Iran and in western Afghanistan and which, as its botanic name implies, has a strong smell and taste of onion.'45 The root of F. alliacea, together with a second alliaceous species of this genus, F. foetida, produces an oil and gum resin commonly called asafoetida. F. alliacea has been identified by Thompson as the source of the Assyrian ;AM,a red-coloured medicinal oil used externally and internally for similar purposes as the reddishhued oil of F. alliacea.'46 During the last century a part of the root of F. alliacea together with the gum it extruded were exported commercially to Bombay from coastal stations in Iran.''47 In the Assyrian period the same product could well have been brought from these Iranian districts, including that of Magan (Makran), to Mesopotamia where the plant does not grow in a natural state. is The Sumerian SUM.SAR MA.GANkI not attested in Assyrian lexical series or medical texts. It is therefore possible to suggest that this plant which had implied alliaceous properties, and which was held in sufficient esteem to be delivered up to the Nanna temple at Ur, may be identified with the important and 'Magan onion' may relate to medicinal root P. alliacea. In this event *AA the same plant.
APPENDIXII

Alexandria Carmania in The reference to Alexander the Great's advance toward Carmnania the main text, introduces the question of the location of Alexandria Carmania. On Ptolemy's map of Persia, it is placed in the south of Carmania. It was clearly not the satrapy capital, CarmaniaMetropolis, which Ptolemy shows farther north in the region of the modern capital, Kirman. Now Alexander spent some days meeting with his satraps, assembled from different parts of Persia, at a place in Carmania which is not otherwise identified by Arrian (Anabasis, vi, xxvii, 3). This gathering occurred while Alexander was moving toward Persepolis so we may expect a fairly southerly location which would have been on the main east to west route extending between Pura (BampfIr)and Persepolis. In following the most direct series of tracks between these two sites, at about half-way to Persepolis, we arrive at the great archaeological mound of Tepe Yahya in
144 See Leemans, Foreign trade, 144, 21. 333-4. 145 Watt, A dictionary of the economic products of India, mI, 146 Thompson, A dictionary of Assyrian botany, 353-9 (iamA ). 147 Watt, op. cit., 330.

A PERIPLUS OF MAGANAND MELUIIHHA

583

southern Kirman/Carmania. The excavator of Tepe has given the level Yahy. of period II there a provisional dating range of 500-300 B.C.148sThis would fall well within the period of Alexander's passage through Carmania. Extensive archaeological reconnaissance undertaken in south Kirman by the Tepe Yahya group, has produced no other major site of this period. It is reasonable to consider, therefore, that the largest and most conveniently placed Achaemenian settlement in the district served as a meeting-place for Alexander and his satraps. Moreover, as towns were not usually identified with Alexander unless he had in some way been associated with them, the suggestion is advanced that the already established Achaemenian settlement at Tepe Yahya could have been called Alexandria Carmania by the later Greek writers (who seem to attest this as a permanent settlement) because of Alexander's proposed stay there. As we have suggested in the main text, a similar association of Alexander with Pura, may have resulted in this town being given the appellation Alexandria Makarene by Greek geographers. Arrian and the other historians of Alexander give neither of these Alexandrias, presumably because the two towns were not new Greek foundations, as were all settlements of this name which the historians do record. As there are extensive Achaemenian and Elamite levels at Tepe Yalhyf, presumably this large site on a major trade route would have been an important settlement in these earlier periods. While it is beyond the scope of the present note to consider the earlier names in detail, it may be useful to recall that Marquarthas associated the southern part of Kirman province with the Achaemenian district of Yutiya.149 If this is a correct identification, perhaps Tepe the largest Achaemenian site of the same region, would seem a reasonable Ya.hya, for the chief settlement of the district. Lastly, we may mention that location many years ago Poebel proposedto identify the region of southern Kirman with the still more ancient Elamite territory Shirikhum.15so This toponym is attested in two texts of Manishtusu of Kish who claims to have conquered the ' King of Anshan and Shirikhum '.151 Several writers have proposed the identification of ancient Anshan with the modern province of Fars in southern Iran.'52 According to this finding, as south Kirman borders on Fars in the east, and, if the listing of Anshan and Shirikhumshould indicate the relative location of these politically allied countries east of Kish, the suggestion of Poebel may prove helpful. The valley in which Tepe Yahya is situated, is called Sughiin. As the loan-words in Persian is irregular, it is not possible to development of Elamnite
148 C. C. Lamberg-Karlovsky, Excavations at Tepe Yahya 1967-1969, Cambridge, Mass., 1970, 5. 149 J. Marquart, A catalogue of the provincial capitals of Erlnshahlr. Marquart arrives at this location by equating Old Persian Yutiya with Middle Persian Krmin Yut, the Yut of Kirmin, and these with the Jut of south Kirmin. 150

A. M. Poebel, University of Pennsylvania. University Museum. Publications of the Baby-

lonian Section, Iv, 1914, 234. 151 Barton, The royal inscriptions, 128, 130. 152 Herzfeld, The Persian Empire, 178-9; Hansman,

Iran, x, 1972, 105.

584

JOHN HANSMAN

draw a clear connexion between this present toponym and Shirikhum. Still, there are important Elamite remains at Tepe YahyA, a site which was unknown to Poebel. Therefore, as with Yutiya in the Achaemenian period, if the district name is correct for the earlier era, it would reasonably follow that the name of the district and of its chief town may have been the same.
ABBREVIATIONS AA American Anthropologist

A f. O An.
APAMNH ARA ARM BASOR BSOAS CAD CAH CHI El EW GJ JCOS JESHO KB KUB LTBA MSL OIP RA RGSP SBE UET ZA

Archiv fir Orientforschung Antiquity

Anthropological Papers of the American Museum of Natural History Ancient records of Assyria and Babylonia (D. D. Luckenbill) Archives royales de Mari Bulletin of the American Schools of Oriental Research Bulletin of the School of Oriental and African Studies Chicago Assyrian dictionary. Cambridge Ancient History Cambridge History of Iran Encyclopaedia of Islam East and West Geographical Journal Journal of Cuneiform Studies Journal of the Economic and Social History of the Orient Keilinschriftliche Bibliothek aus Boghazkii Keilschrifturkunden Die lexikalischen Tafelserien der Babylonier und Assyrer in den Berliner Museen Materialien zum sumerischen Lexikon University of Chicago. Oriental Institute Publications Revue d'Assyriologie Records of the Geological Survey of Pakistan Sacred Books of the East Ur Excavation Texts Zeitschrift fiUr Assyriologie ANNEXE BY

H. W.

BAILEY

Mleccha-, Bal5, and Gadrasia A. mlecchd-; verbal mlhcchati,mlistd-, mlecchitaI. 1. Earliest reference is in the later Veda, &atapathabrahmana,3.2.1.24: the noun mlecchd-,used of Asura celestial beings who speak imprecise language whether ill-pronounced or foreign. The word helayo, variant hailo, is quoted. No localization is given for this mythic allusion. 2. Epic usage. Mahabharatacontrasts the mleccha-with the arya-, and has ' the mleccha-bhasa Mlecchalanguage ', and mleccha-vik' using Mlecchaspeech '. The Dharmasitra text, the Manu-smrti, 2.23, has the mleccha-desd-' Mleccha country' as unfit for Brahmanical sacrifices. II. Localization. 1. The Mahabharataplaces Mleccha loosely in east, north, and west. The has Mleccha for the Matsya people of Rajputana (see S. Levi, Journal Asiatique, XIe S~r., xI, 1, 1918, 123). Ramaya.a

A PERIPLUS

OF MAGAN AND MELUHUIA

585

2. Varahamihira, c. A.D. 550, placed the Mleccha in the upara- region, the western. His upara- region refers to the peoples beyond the Sindhu, Indus, for whom the Mahabhdratahad the epithet pdre-sindhavah'beyond the Sindhu'. Var5hamihira has peoples reaching from Vokkana- 'Wakhan ', through Pancanada- 'Panjab ', to the Parata-, Parada-, which is the Greek IIapact1v4 'Parada-' placed by Ptolemy in Gedrosia. These Parada- are named in the Paikuli inscription of the Sasanians and in the inscription of Shahpuhr I, Parthian text, line 2, in the list krmn skstn twgrnmkwrnp'rtn hndstn ' Kirman, Sakastan, Tugran, Partan, Hindastan '. This position excludes LBvi's proposal of the Panjab for the Parata-. These Indian localizations give only 'beyond the Indus '. III. Linguistic evidence 1. (a) Later Veda, mlecchd- and verbal mlicchati, with participle in the Scholiast to Panini mlista-; mlecchita-is also cited. Patanjali has the infinitive mlecchitavai. (b) Pall, in the oldest texts Digha-niklya and Vinya, milakkhu-, milakkhuka-, milakkha-, milakkha-bhJsa,and later milica-. (c) Jaina older Ardha-magadhi, milikkha- (with vokkana- and yavana'Wakhan' and 'Greek '), milakkhu-, milikkhu-, mileccha-, and Maharastri milittha- ' speaking indistinctly '. (d) Buddhist Sanskrit mlecha-, whence Saka Khotan milaicha-. (e) New Indo-Aryan in R. L. Turner, Comparative dictionary, no. 10398, Kasmiri mih (with -6h from older -cch-, not -ks-); Bengali mech of a TibetoBurmese tribe, Sinhalese milidu, miliiidu 'savage', milis, maladu, Panjabi milech, malech. The Pali -kkh- was explained as secondary to -cch- by J. Wackernagel, Altindische Grammatilc,I, 154; but was unexplained according to Turner, loc. cit. 2. The starting-point of the interpretation should be a form *mleks-,mliks-. Within the Veda there is variation between -cch- (-ch-) and -ks- as in Atharvaveda beside Gukla-yajur-veda, Vajasaneyi-samhita 'fetter', rcchdr5.rks6iland within the Atharva-veda in pariksit- and variant paricchit- ' surrounding '. Hence datapathabrahmanamlecchd-may be traced to older *mleksa-. The ks was replaced by -kkh- or by retroflex -ch- or by palatalized -cch- in different dialects. Within the Veda there was also variation in kId-, ksa-, and khyd-from kad-, correspondingto Avestan xsd- from kas- 'look at'. If the oldest form had then *mleksa-,this -ks- could be accepted as a substitute for a foreign velar fricative -XX-(the sound expressed in Arabic script by

t kh).
If the word *mleks- was a foreign name, it was adapted to the usual Vedic verbal system, giving participle in the grammarians, supported by the mlistaJaina Maharastrimilittha-. The vowel -e- of mlecchd-was thus adapted into the ablaut system -e-: -i-.
VOL. XXXVI. PART 3. 40

586

JOHN HANSMLAN

For recent comments on mleccha,see Wackernagel, Altindische Grammatik. Introduction gindrale. Nouvelle 4dition.. . par Louis Renou, 1957, 73; M. Mayrhofer, Kurzgefasstes etymologisches Warterbuchdes Altindischen, 699, mleccha. B. Bal55 1. The name is recorded first in Arabic blws (with the normal early Arabic use of s (sad) for Persian 6), and Persian blwj in the Hudid al-'dlam (A.D. 982) with j (~im)for the Persian Ewhich was at first not distinguished. The vocalization of the Arabic forms is reported (P. Schwartz, Iran im Mittelalter,I, 260-1) as balls and bules, billis, these two replacements of Persian bal-. 2. Earliest localization in al-Mas'iidi (A.D. 943), al-Muqaddasi (A.D. 985), and IHudildal-'dlam (A.D. 982): Bal56 in the desert, Kafi5in the mountains. Al-Tabari, enumerating the enemies of AnbSaruvandoes not mention the name of Bal15, hence the poetic Shihnahma c. A.D. 1020 has replaced another name by Bala6, possibly the North Caucasian name blnjr, *balanlar (?), of a people living north of Darband. Arabic historians associate the Baliis with the Qufs (Persian Kafi6 ' mountain men ') north of Kirman. 3. The modern Bal15i language is distinct from the south-western dialect of Fars, though some loan-words are found thence, and it has features like the northern dialect, as in Parthian (note Bal15i zirdd, Parthian of Turfan texts zyrd, *zirs, 'heart '). Two modern Iranian dialects still spoken in Afghanistan, the Ormuri (or Baraki) and Para6i, show non-eastern dialect types and may be survivals of the pre-Paito (Saka) language group. 4. The word balu6is reported in south-eastern Iran as being used with the meaning 'shepherd'. Dr. I. Gershevitch has informed me that the Baikirdi language used both Bal65 and Kurd to mean 'shepherd '. This would seem to show that ethnic names have been used in a non-ethnic sense. 5. The name Bal15, being attested only during the Islamic period, may have either an Iranian or a non-Iranian origin. (a) Assuming an Iranian origin from some non-eastern Iranian dialect (if the word belonged originally to the Bal16i language) or from an eastern dialect (if it has entered secondarily into Bal16i), the older forms of the word possible by regular change are numerous. The initial could be b-, or v-, or m-, the medial consonant could be -1-, -r-, -rd-, -rt-, or -dr-. The syllable -66 could be either older -aun- or -aut.5-. If the word is transmitted irregularly there is no control. is (b) If the name Bal655 non-Iranian, it would be necessary to show how the now speaking Iranian, had received the name. (See above, p. 569.) Bal15, C. Gadrasia The ancient name of modern eastern Bal5istln (now in Pakistan) was in Greek transmission Gadrasia, Gedrasia, "E8pcooa with the ethnic Faspoitta, name E8ptdCToL and from the eparchy FE Gedr5sanoi. pwarvol,

A PERIPLUS OF MAGAN AND MELUHIHA

587

The adjacent province to the north was in Old Persian spelling harahvati (h r u vt i), easily recognizable as formed from harah-, Old Indian saras-' watery land', Greek "Aos'swamp', and formed like the Old Indian sarasvati-, the This name was recorded in Greek name of a river from the time of the .Rgveda. Arakh tos (river name), Arakh toi (the inhabitants), and Arakhasia (the land). A similar replacement of -avati occurs in Greek Hudra~tas for the Old Indian river of the Panjab name Irivati. This identity of second components is important here for the interpretation of Gadrasia by a Persian word ending in -avati 'the land possessing the feature expressed by gadra-'. A feature which struck the Greek geographers was that Gadrasia had abundant water from hudreia 'underground water'. The word gadra-would thus suit well to represent this concept of' hole, well ' or' undergroundchannel ', the well-known system used in Iran to irrigate. This is the system called in Zoroastrian Pahlavi texts kts, Persian of Turfan khs, *ka~as, kahas, 'channel', and in New Persian kahre'z, karez, Bal15i karez. This word kahrezis fairly surely from an older *kanOra-'digging, hole, excavation' with the familiar suffix -e~ (or -i6). For this system of 'channel water' the Avesta had the word vadi- which is found in modern dialects, Palto wila from w~la, Shuglhniwe'8. *vadj- from vadi- or vadi-, Wakhi wad, Sarikoli wa8, Yidghha Variation of suffix is common, hence a derivative *vadra-is likely. From at least the time of the Indo-Parthian king Gondophares the initial Iranian vi- is attested as gu-, in the on coins Guduphara-, Kharosthi spelling beside Greek Undopheros, from older vinda(t)-farnah-. This is a change which was carried through for all initials v- in Para5i, Ormuri,and Biyabanaki, and similarly to gw- in Bal15i, which G. Morgenstierne thought might be due to a substrate language at a time after the Bal5I speakers met Indo-Aryan speakers (IIFL, I, 317). These various points would show *uadra- 'underground channel' replaced by *gadra-, and used to form the province name *gadravati,preserved in Greek transmission as Gadrbsia. As one of many possible origins for an Iranian word Bal15, this *uadravati could have resulted in *ba8laut-and thence, with adjectival suffix -E,in balS3in a dialect which replaced initial v- by b-, as first attested in the ninth century. To be noted also is the development of the old name Arakhdsia, Harahvati in Greek Khorokhoad, Xopoxoa8, and in Arabic rx3, *rukhadh, and al-rukhaj, that is, ArokhaE,with the final adjectival -. Both the suffix -c, New Persian -z, and -cik, New Persian -zi, are well attested in adjectives for ethnic names. To connect Gadrasia with Bal55 in this way is purely speculative, except that it attempts to explain the second part of the name. The -5c of the word must also receive its full weight. Bal655