The Ancient Geography of India/Western India

From Jatland Wiki
Jump to navigation Jump to search
Wikified by:Laxman Burdak, IFS (Retd.)
Back to Index of the Book
The Ancient Geography of India: I.
The Buddhist Period, Including the Campaigns of Alexander, and the Travels of Hwen-Thsang
Sir Alexander Cunningham
Trübner and Company, 1871 - India

Western India

[p.248]: Western India, according to Hwen Thsang, was divided into three great states, named Sindh, Gurjjara, and Balabhi. The first comprised the whole valley of the Indus from the Punjab to the sea, including the Delta and the island of Kachh ; the second comprised Western Rajasthan and the Indian Desert, and the third comprised the peninsula of Gujarat, with a small portion of the adjacent coast.

I. Sindh.

In the seventh century Sindh was divided into four principalities, which, for the sake of greater distinctness, I will describe by their geographical positions, as 1. Upper Sindh, 2. Middle Sindh, 3. Lower Sindh, and 4. Kachh. 1

The whole formed one kingdom under the Raja of Upper Sindh, who, at the time of Hwen Thsang's visit in A.D. 641, was a Siu-to-lo or Sudra. So also in the time of Chach, only a few years later, the minister Budhiman informs the king that the country had been formerly divided into four districts, each under its own ruler, who acknowledged the supremacy of Chach's predecessors. 2 At a still earlier date Sindh is said to have been divided into four principalities by Ayand, the son of Kafand 3 who reigned some time after Alexander the Great. These four principalities are named Zor, Askalandusa, Samid, and

1 See Map No. IX.

2 Postans in Journ. Asiat. Soc. Bengal, 1838, p. 93.

3 Rashid ud din, in Reinaud's 'Fragments Arabes,' p. 47.

[p. 249]: Lohana, all of which will be discussed presently, as they would appear to correspond with the divisions noted by Hwen Thsang.

Upper Sindh.

The single principality of Upper Sindh, which is now generally known as Siro, that is the "Head or Upper" division, is described as being 7000 li, or 1167 miles, in circuit, which is too great, unless, as is very probable, it comprised the whole of Kachh Gandava on the west. This was, no doubt, always the case under a strong government, which that of Chach's predecessor is known to have been. Under this view Upper Sindh would have comprised the present districts of Kachh-Gandava, Kahan, Shikarpur, and Larkana to the west of the Indus, and to the east those of Sabzalkot and Khairpur. The lengths of the frontier lines would, therefore, have been as follows : — on the north 340 miles ; on the west 250 miles ; on the east 280 miles, and on the south 260 miles; or altogether 1030 miles, which is a very near approximation to the estimate of Hwen Thsang.

In the seventh century the capital of the province was named Pi-chen-po-pu-lo, which M. Julien transcribes as Vichava-pura. M. Vivien de St. Martin, however, suggests that it may be the Sanskrit Vichala-pura, or city of " Middle Sindh," which is called Vicholo by the people. But the Sindhi and Panjabi Vich and the Hindi Bich, or " middle," are not derived from the Sanskrit, which has a radical word of its own, Madhya, to express the same thing. If Hwen Thsang had used the vernacular terms, his name might have been rendered exactly by the Hindi Bichwa-pur, or

[p. 250]: "Middle City," but as he invariably uses the Sanskrit forms, I think that we must rather look to some pure Sanskrit word for the original of his Pi-chen-po-pu-lo. Now we know from tradition, as well as from the native historians, that Alor was the capital of Sindh both before and after the period of Hwen Thsang's visit; this new name, therefore, must be only some variant appellation of the old city, and not that of a second capital. During the Hindu period it was the custom to give several names to all the larger cities, — as we have already seen in the case of Multan. Some of these were only poetical epithets ; as Kusuma-pura, or " Flower City " applied to Pataliputra, and Padmavati, or, "Lotus Town" applied to Narwar ; others were descriptive epithets as Varanasi, or Banaras, applied to the city of Kasi, to show that it was situated between the Varana and Asi rivulets ; and Kanyakubja, the " hump-backed maiden," applied to Kanoj, as the scene of a well-known legend. The difference of name does not, therefore, imply a new capital, as it may be only a new appellation of the old city, or perhaps even the restoration of an old name which had been temporarily supplanted. It is true that no second name of Alor is mentioned by the historians of Sindh ; but as Alor was actually the capital in the time of Hwen Thsang, it would seem to be quite certain that his name of Pi-chen-po-pu-lo is only another name for that city.

It is of importance that this identification should be clearly established, as the pilgrim places the capital to the west of the Indus, whereas the present ruins of Alor or Aror are to the east of the river. But this very difference confirms the accuracy of the identifi-

[p. 251]: cation, for the Indus formerly flowed to the east of Alor, down the old channel, now called Nara, and the change in its course did not take place until the reign of Raja Dahir, 1 or about fifty years after Hwen Thsang's visit. The native histories attribute the desertion of Alor by the Indus to the wickedness of Raja Dahir ; but the gradual westing of all the Panjab rivers which flow from north to south, is only the natural result of the earth's continued revolution from west to east, which gives their waters a permanent bias towards the western banks, 2 The original course of the Indus was to the east of the Alor range of hills ; but as the waters gradually worked their way to the westward, they at last turned the northern end of the range at Rori, and cut a passage for themselves through the gap in the limestone rocks between Rori and Bhakar. As the change is assigned to the beginning of Dahir' s reign, it must have taken place shortly after his accession in A.D. 680 ; — and as Muhammad Kasim, just thirty years later, was obliged to cross the Indus to reach Alor, it is certain that the river was permanently fixed in its present channel before A.D. 711.

The old bed of the Indus still exists under the name of Nara, and its course has been surveyed from the ruins of Alor to the Ran of Kachh. From Alor to Jakrao, a distance of 100 miles, its direction is nearly due south. It there divides into several channels, each bearing a separate name. The most easterly

1 Postans, Journ. Asiat. Soc. Bengal, 1838, p. 103.

2 All streams that flow from the poles towards the equator work gradually to the westward, while those that flow from the equator towards the poles work gradually to the eastward. These opposite effects are caused by the same difference of the earth's polar and equatorial velocities which gives rise to the trade winds.

[p. 252]: channel, which retains the name of Nara, runs to the south-east by Kipra and Umarkot, near which it turns to the south-west by Wanga Bazar and Romaka Bazar, and is there lost in the great Ran of Kachh.

The most westerly channel, which is named Purana, or the " Old River," flows to the south-south-west, past the ruins of Brahmanabad and Nasirpur to Haidarabad, below which it divides into two branches. Of these, one turns to the south-west and falls into the present river 15 miles below Haidarabad and 12 miles above Jarak. The other, called the Guni, turns to the south-east and joins the Nara above Romaka Bazar.

There are at least two other channels between the Purana and the Nara, which branch off just below Jakrao, but their courses arc only partially known.

The upper half of the old Nara, from Alor to Jakrao, is a dry sandy bed, which is occasionally filled by the flood waters of the Indus. From its head down to Jamiji it is bounded on the west by a continuation of the Alor hills, and is generally from 200 feet to 300 feet wide and 20 feet deep. From Jamiji to Jakrao, where the channel widens to 600 feet with the depth of 12 feet, the Nara is bounded on both sides by broad ranges of low sand-hills. Below Jakrao the sand-hills on the western bank suddenly terminate, and the Nara, spreading over the alluvial plains, is divided into two main branches, which grow wider and shallower as they advance, until the western channels are lost in the hard plain, and the eastern channels in a succession of marshes. But they reappear once more below the parallel of Hala and Kipra, and continue their courses as already described above. 1

1 See Map No. IX.

[p. 253]: In Upper Sindh the only places of ancient note are Alor, Rori-Bhakar, and Mahorta, near Larkana. Several other places are mentioned in the campaigns of Alexander, Chach, Muhammad bin Kasim, and Husen Shah Arghun ; but as the distances are rarely given, it is difficult to identify the positions where names are so constantly changed. In the campaign of Alexander we have the names of the Massance, the Sogdi, the Musikani, and the Prasti, all of which must certainly be looked for in Upper Sindh, and which I will now attempt to identify.

Massanae and Sodrae, or Sogdi.

On leaving the confluence of the Panjab rivers, Alexander sailed down the Indus to the realm of the Sogdi, Sodrae, where, according to Arrian, 1 " he built another city." Diodorus2 describes the same people, but under a different name : — "Continuing his descent (if the river, he received the submission of the Sodrae and the Mussana, nations on opposite banks of the stream, and founded another Alexandria, in which he placed 10,000 inhabitants." The same people are described by Curtius, 3 although he does not mention their names : — " On the fourth day he came to other nations, where he built a town called Alexandria." From these accounts it is evident that the Sogdi of Arrian and the Sodra of Diodorus are the same people, although the former have been identified with the Sodha Rajputs by Tod and M'Murdo, the latter with the servile Sudras by Mr. Vaux. The Sodhas, who are a branch of the Pramaras, now occupy the south-

1 ' Anabasis,' vi. 15.

2 Hist. Univers. xvii. 56.

3 Vita Alex., ix. 8.

[p. 254]: eastern district of Sindh, about Umarkot, but according to M'Murdo, 1 who is generally a most trustworthy guide, there is good reason to believe that they once held large possessions on the banks of the Indus, to the northward of Alor. In adopting this extension of the territory formerly held by the Sodha Rajputs, I am partly influenced by the statement of Abul Fazl, that the country from Bhakar to Umarkot was peopled by the Sodas and Jharejas in the time of Akbar, 2 and partly by the belief that the Massana of Diodorus are the Musarnei of Ptolemy, whose name still exists in the district of Muzarka, to the west of the Indus below Mithankot. Ptolemy also gives a town called Musarna, which he places on a small affluent of the Indus, to the north of the Askana rivulet. The MmantK affluent may therefore be the rivulet of Kahan, which flows past Pulaji and Shahpur, towards Khangarha or Jaeobabad, and Musarna may be the town of Shahpur, which was a place of some consequence before the rise of Shikarpur. " The neighbouring country, now nearly desolate, has traces of cultivation to a considerable extent. " The Sogdi, or Sodrae, I would identify with the people of Seorai, which was captured by Husen Shah Arghun on his way from Bhakar to Multan. 4 In his time, A.D. 1525, it is described as "the strongest fort in that country." It was, however, deserted by the garrison, and the conqueror ordered its walls to be razed to the ground. Its actual position is unknown, but it was

1 Journ. Royal Asiat. Soc, i. 33. i ' Ayin Akbari,' ii. 117.

2 Thornlon, ' Gazetteer,' in voce.

3 Erskine's Hist, of India, i. 388. Postans, Journ. Asiat. Soc. Bengal, 1811, 275.

[p. 255]: probably close to Fazilpur, halfway between Sabzalkot and Chota Ahmedpur, where Masson 1 heard that there was formerly a considerable town, and that "the wells belonging to it, 360 in number, were still to be seen in the jangals." Now in this very position, that is about 8 miles to the north-east of Sabzalkot, the old maps insert a village named Sirwahi, which may possibly represent the Seorai of Sindhian history. It is 96 miles in a direct line below Uchh, and 85 miles above Alor, or very nearly midway between them. By water the distance from Uchh would be at least one-third greater, or not less than 120 miles, which would agree with the statement of Curtius that Alexander reached the place on the fourth day. It is admitted that these identifications are not altogether satisfactory ; but they are perhaps as precise as can now be made, when we consider the numerous fluctuations of the Indus, and the repeated changes of the names of places on its banks. One fact, preserved by Arrian, is strongly in favour of the identification of the old site near Fazilpur with the town of the Sogdi, namely, that from this point Alexander dispatched Kraterus2 with the main body of the army, and all the elephants, through the confines of the Arachoti and Drangi. Now the most frequented Ghat for the crossing of the Indus towards the west, via the Gandava and Bolan Pass, lies between Fazilpur on the left bank, and Kasmor on the right bank. And as the ghats, or points of passage of the rivers, always determine the roads, I infer that Kraterus must have begun his long march towards Arachosia and Drangiana from this place, which is the most northern

1 ' Travels,' i. 382. 2 'Anabasis,' vi. 15.

[p. 256]: position on the Indus for the departure of a large army to the westward. It seems probable, however, that Kraterus was detained for some time by the revolt of Musikanus, as his departure is again mentioned by Arrian, 1 after Alexander's capture of the Brahman city near Sindomana.

Between Multan and Alor the native historians, as well as the early Arab geographers, place a strong fort named Bhatia, which, from its position, has a good claim to be identified with the city which Alexander built amongst the Sogdi, as it is not likely that there were many advantageous sites in this level tract of country. Unfortunately, the name is variously written by the different authorities. Thus, Postans gives Paya, Bahiya, and Pahiya ; Sir Henry Elliot gives Pabiya, Batia, and Bhatiya, while Price gives Bahatia. 2 It seems probable that it is the same place as Talhati, 3 where Jam Janar crossed the Indus ; and perhaps also the same as Matila, or Mahatila, 4 which was one of the six great forts of Sindh in the seventh century.

Bhatia is described by Ferishta as a very strong place, defended by a lofty wall and a deep broad ditch. 5 It was taken by assault in a.h. 393, or A.D. 1003, by Mahmud of Ghazni, after an obstinate defence, in which the Raja, named Bajjar or Bije Rai, was killed. Amongst the plunder Mahmud obtained no less than 280 elephants, a most substantial proof of the wealth and power of the Hindu prince.

1 ' Anabasis,' vi. 17.

2 Dowson's edition of Sir H. Elliot, i. 138.

3 Journ. Asiat. Soc. Bengal, 1845, p. 171.

4 Ibid., 184,5, p. 79.

5 Briggs's ' Ferishta,' i. 39 ; and Tabakat.i.Akbari, in [[Sir Henry Elliot]], p. 186.


[p. 257]: From the territory of the Sogdi or Sodrae, Alexander continued his voyage down the Indus to the capital of a king named Musikanus, according to Strabo, Diodorus, and Arrian, 1 or of a people named Musicani, according to Curtius. 2 From Arrian we learn that this kingdom had been described to Alexander as " the richest and most populous throughout all India ; " and from Strabo we get the account of Onesikritus that " the country produced everything in abundance ; " which shows that the Greeks themselves must have been struck with its fertility. Now these statements can apply only to the rich and powerful kingdom of Upper Sindh, of which Alor is known to have been the capital for many ages. Where distances are not given, and names disagree, it is difficult to determine the position of any place from a general description, unless there are some peculiarities of site or construction, or other properties which may serve to fix its identity. In the present instance we have nothing to guide us but the general description that the kingdom of Musikanus was " the richest and most populous throughout all India." But as the native histories and traditions of Sindh agree in stating that Alor was the ancient metropolis of the country, it seems almost certain that it must be the capital of Musikanus, otherwise this famous city would be altogether unnoticed by Alexander's historians, which is highly improbable, if not quite

1 Strabo, Geogr., xv. i. 22-34 and 54. Diodorus, xvii. 10. Arrian, ' Anabasis,' vi. 15.

2 Vita Alex., ix. 8.

[p. 258]: impossible. That the territory of Alor was rich and fertile we know from the early Arab geographers, who are unanimous in its praise.

The ruins of Alor are situated to the south of a gap in the low range of limestone-hills, which stretches from Bhakar towards the south for about 20 miles, until it is lost in the broad belt of sand-hills which bound the Nara, or old bed of the Indus, on the west. Through this gap a branch of the Indus once flowed, which protected the city on the north-west. To the north-east it was covered by a second branch of the river, which flowed nearly at right angles to the other, at a distance of 3 miles. At the accession of Raja Dahir, in A.D. 680, the latter was probably the main stream of the Indus, which had been gradually working to the westward from its original bed in the old Nara. 1 According to the native histories, the final change was hastened by the excavation of a channel through the northern end of the range of hills between Bhakar and Rori.

The true name of Alor is not quite certain. The common pronunciation at present is Aror, but it seems probable that the original name was Rora, and that the initial vowel was derived from the Arabic prefix Al, as it is written Alror in Biladuri, Edrisi, and other Arab authors. This derivation is countenanced by the name of the neighbouring town of Rori, as it is a common practice in India thus to duplicate names. So Rora and Rori would mean Great and Little Rora. This word has no meaning in Sanskrit, but in Hindi it signifies "noise, clamour, roar" and also "fame." It is just possible, therefore, that the full name of the

1 See Map No. IX.

[p. 259]: city may have been Rora-pura, or Rora-nagara, the " Famous City." This signification suggested itself to me on seeing the name of Abhijanu applied to a neighbouring village at the foot of the hill, 2 miles to the south-west of the ruins of Alor. Abhijan is a Sanskrit term for " fame," and is not improbably connected with Hwen Thsang's Pi-chen-po-pu-lo, which, by adding an initial syllable o, might be read as Abhjanwapura. I think it probable that Alor may be the Binagara of Ptolemy, as it is placed on the Indus to the eastward of Oskana, which appears to be the Oxykanus of Arrian and Curtius. Ptolemy's name of Binagara is perhaps only a variant reading of the Chinese form, as pulo, or pura, is the same as nagara, and Pichenpo may be the full form of the initial syllable Bi.

The city of Musikanus was evidently a position of some consequence, as Arrian relates that Alexander " ordered Kraterus to build a castle in the city, and himself tarried there to see it finished. This done, he left a strong garrison therein, because this fort seemed extremely commodious for bridling the neighbouring nations and keeping them in subjection." It was no doubt for this very reason that Alor was originally founded, and that it continued to be occupied until deserted by the river, when it was supplanted by the strong fort of Bhakar.

PraestiPortikanus, or Oxykanus.

From the capital of Musikanus Alexander allowed his fleet of boats to continue their course down the Indus, while he himself, according to Arrian, 1

1 ' Anabasis,' vi. 16.

[p. 260]: marched against a neighbouring prince named Oxykanus, and took two of his chief cities at the first assault. Curtius makes Oxykanus the king of a people named Praesti1 and states that Alexander captured his chief city after a siege of three days. Diodorus and Strabo call the king Portikanus. Now, these various readings at once suggest the probability that the name was that of the city, which, either as Uchcha-gam, or Porta-gam, means simply the "Lofty town," in allusion to its height. The description of Curtius of the " tremendous crash" made by the fall of two towers of its citadel shows that the place must have been more than usually lofty. I would therefore identify it with the great mound of Mahorta on the bank of the Ghar river, 10 miles from Larkana. Masson describes it as "the remains of an ancient fortress, on a huge mound, named Maihota 2 . Mahorta, which is the spelling adopted by the surveyors, is probably Mahorddha, for maha+urddha+grama, or "the great lofty city," which, as pure Sanskrit, is not likely to be a modern name. This identification appears to me to be very probable, not only on account of the exact correspondence of name, but also on account of the relative positions of Alor and Mahorta with reference to the old course of the Indus. At present Mahorta is within a few miles of the river ; but in the time of Alexander, when the Indus flowed down the bed of the Nara, the nearest point of the stream was at Alor, from which Mahorta was distant 45 miles to the south of west. Hence Alexander was obliged to leave his fleet, and to march against Oxykanus

1 Vita Alex., ix. 8, 26. 2 'Travels,' i. 461.

[p. 261]: The site of Mahorta must always have been a position of great importance, both commercially and politically, as it commanded the high-road from Sindh, via, Kachh-Gandava, to Kandahar. Since its desertion, the same advantages have made Larkana, which is situated on the same small stream, 10 miles to the west of Mahorta, one of the most flourishing places in Sindh. The rivulet called the Ghar rises near Kelat, and traverses the whole length of the Mula, or Gandava Pass, below which it is now lost in the desert. But the channel is still traceable, and the stream reappears on the frontier of Sindh, and flows past Larkana and Mahorta into the Indus. Under a strong and judicious ruler, who could enforce an economical distribution of the available waters, the banks of the Ghar rivulet must formerly have been one of the most fertile districts of Sindh.

The name of Praesti given by Curtius 1 might, according to Wilson, be applied to a people occupying the thals, or " oases," of the desert. He refers to Prastha, or Prasthala, as derived from sthala, the Sanskrit form of the vernacular thal, which is the term generally used to designate any oasis in Western India. But as the name is simply Praesti, I think that it may rather be referred to prastha, which means any clear piece of level ground, and might therefore be applied to the plain country about Larkana, in contradistinction to the neighbouring hilly districts of Sehwan and Gandava. It seems possible, however, that it may be connected with the Piska of Ptolemy, which he places on the lower course of the small stream that flows past Oskana into the Indus.

1 Vita Alex., ix. 8.

[p. 262]: Now Oskana is almost certainly the Oxykanus of Arrian and Curtius, for not only are the two names absolutely identical, but the inland position of Oskana, on a small stream to the west of the Indus, agrees exactly with that of Mahorta, which I have identified with Oxykanus. I think also that Ptolemy's Badana, which lies immediately to the north of the rivulet, must be the present Gandava, as the letters B and G are constantly interchanged. In the books of the early Arab writers it is always called Kandabil.

2. Middle Sindh

The principality of Middle Sindh, which is generally known as Vichnlo, or the " Midland," is described by Hwen Thsang as only 2500 li, or 417 miles, in circuit. With these small dimensions the province must have been limited to the modern district of Sehwan, with the northern parts of Haidarabad and Umarkot. Within these limits the north and south frontiers are each about 160 miles in length, and the east and west frontiers about 45 miles each, or altogether not more than 410 miles in circuit. The chief city, named, was situated at 700 li, or 117 miles, from the capital of Upper Sindh, and 50 miles from Pitasila, the capital of Lower Sindh. As the former was Alor, and the latter was almost certainly the Pattala of the Greeks, or Haidarabad, the recorded distances fix the position of O-fan-cha in the immediate vicinity of the ruins of an ancient city called Bambhra-ka-Tul, or the " Ruined Tower," or simply Banbhar, which, according to tradition, was the site of the once famous city of Brahmanwas, or Brahmanabad. Hwen Thsang's kingdom of Ofancha, or Avanda, therefore, corresponds

[p. 263]:

as nearly as possible with the province of middle Sindh, which is now called Vichalo.

At the present day the principal places in this division of Sindh are Sehwan, Hala, Haidarabad, and Umarkot. In the middle ages, under Hindu rule, the great cities were Sadusan, Brahmana, or Bahmanwa, and Nirunkot. But as I shall presently attempt to show that Nirunkot was most probably the modern Haidarabad and the ancient Pattala, it will more properly be included in the province of Lower Sindh, or Lar. Close to Bahmanwa the early Muhammadans founded Mansura, which, as the residence of their governors, was the actual capital of the province, and soon became the largest city in all Sindh. In the time of Alexander, the only places mentioned are Sindomana, and a city of Brahmans, named Harma-telia by Diodorus. I will now describe these places in detail, beginning with the most northerly.

Sindomana, or Sehwan.

[p. 263]: From the city of Oxykanus, Alexander " led his forces against Sambus, whom he had before declared governor of the Indian mountaineers." The Raja abandoned his capital, named Sindomana, which, according to Arrian, 1 was delivered up to Alexander by the friends and domestics of Sambus, who came forth to meet him with presents of money and elephants. Curtius 2 calls the raja Sabus, but does not name his capital. He simply states that Alexander, having received the " submission of several towns, captured the strongest by mining." The narrative of Diodorus.3

1 'Anabasis,' vi. 16.

2 Vita Alex., is. 8.

3 Hist. Univers., xvii. 50.

[p. 264]:also omits the name of the capital, but states that Sambus retired to a great distance with thirty elephants. Strabo1 merely mentions Raja Sabus, and Sindomana his capital, without adding any particulars. Curtius2 alone notes that Alexander returned to his fleet after the capture of the raja's strongest city, which must therefore have been at some distance from the Indus.

I agree with all previous writers on the ancient geography of this part of India in identifying Sindomana with Sehwan ; partly from its similarity of name, and partly from its vicinity to the Lakki mountains. Of its antiquity there can be no doubt, as the great mound, which was once the citadel, is formed chiefly of ruined buildings, the accumulation of ages, on a scarped rock, at the end of the Lakki range of hills. De La Hoste3 describes it as an oval, 1200 feet long, 750 feet broad, and 80 feet high ; but when I saw it in 1855, it appeared to me to be almost square in shape, and I judged it to be somewhat larger and and rather more lofty above the river bed than Burnes's estimate. 4 It was then on the main stream of the Indus; but the river is constantly changing its channel, and in all the old maps it is placed on a western branch of the Indus. In ancient times, however, when the river flowed down the eastern channel of the Nara, Sehwan was not less than 65 miles distant from its nearest

1 Geogr., xv. 1, 32.

2 Vita Alex., ix. 8 : " Rursus amnem, in quo classem expectare se jusserat, repetit."

3 Journ. Asiat. Soc. Bengal, 1840, p. 913.

4 Westmacott, in Journ. Asiat. Soc. Bengal for 1840, p. 1209, says about 100 feet above the Arrul river, or Aral, which, in his time was a branch of the Indus.

[p. 265]:point at Jakrao, where it leaves the sand-hills. At present its water supply is entirely derived from the Indus, which not only flows under the eastern front of the town, but also along its northern front, by a channel called the Aral river, from the great Manchur lake, which is supplied by the other Nara, or great western branch of the Indus. But as the site could not have been occupied unless well supplied with water, it is certain that the Manchur lake must have existed long previous to the change in the course of the Indus. Judging by its great depth in the middle, 1 it must be a natural depression ; and as it is still fed by two small streams, which take their rise in the Hala Lakki mountains, to the south, it seems probable that the lake may have extended even up to the walls of Sehwan, before the floods of the western Nara cut a channel into the Indus, and thus permanently lowered the level of its waters. The lake abounds in fish, from which it would appear to derive its name, as Manchur is but a slight alteration of the Sanskrit Matsya, and the Hindi machh, or machhi, " fish." I think, therefore, that Manchar may be only a familiar contraction of machhi-wala Tal, or Fish Lake.

The favourable position of Sehwan, on a lofty isolated rock, near a large lake, with food and water in abundance, would certainly have attracted the notice of the first inhabitants of Sindh. We find, accordingly, that its early occupation is admitted by all inquirers. Thus, M'Murdo2 says, " Sehwan is undoubtedly a place of vast antiquity ; perhaps more so than either Alor or Bahmana." The present name is

1 Westmacott, Journ. Asiat. Soc. Bengal, 1810, p. 1207.

2 Journ. Royal Asiat. Soc, i. 30.

[p. 266]:said to be a contraction of Sewistan, which was so called after its inhabitants, the Sewis, or Sabis. But in all the early Arab geographers the name is somewhat differently written, as Sadustan, or Sadusan, or Sarusan, of which the first two syllables agree with the Greek Sindomana. I therefore reject the reading of Sewistan as a modern innovation of the Hindus, to connect the place with the name of their god Siva. The Sindo of the Greek, and the Sadu of the early Muhammadans, point to the Sanskrit name of the country, Sindhu, or to that of its inhabitants, Saindhava, or Saindhu, as it is usually pronounced. Their stronghold, or capital, would therefore have been called Saindhava-sthana, or Saindhu-sthan, which, by the elision of the nasal, becomes the Sadustan of the Arab geographers. In a similar manner Wilson derives the Greek Sindomana from " a very allowable Sanskrit compound, Sindu-man, the "possessor of Sindh." I am inclined, however to refer the Greek name to Saindhava-vanam , or Saindhuwan, as the " abode of the Saindhavas."

It seems strange that a notable place like Sehwan should not be mentioned by Ptolemy under any recognizable name. If we take Haidarabad as the most probable head of the Delta in ancient times, then Ptolemy's Sydros, which is on the eastern bank of the Indus, may perhaps be identified with the old site of Mattali, 12 miles above Haidarabad, and his Pasipeda with Sehwan. The identification of Ptolemy's Oskana with the Oxykanus, or Portikanus, of Alexander, and with the great mound of Mahorta of the present day, is, I think, almost certain. If so, either Piska or Pasipeda must be Sehwan.

[p. 267]: Hwen Thsang takes no notice of Sehwan, but it is mentioned in the native histories of Sindh as one of the towns captured by Muhammad bin Kasim in A.D. 711. It was again captured by Mahmud of Ghazni in the beginning of the eleventh century ; and under the Muhammadan rule it would appear to have become one of the most flourishing places in Sindh. It is now very much decayed, but its position is so favourable that it is not likely ever to be deserted.

Brahmana or Brahmanabad

[p. 267]:From Sindomana Alexander " marched back to the river, where he had ordered his fleet to wait for him. Thence, descending the stream, he came on the fourth day to a town through which was a road to the kingdom of Sabus." 1 When Alexander quitted his fleet at Alor (the capital of Musikanus) to march against Oxykanus, he had no intention of going to Sindomana, as Raja Sambus, having tendered his submission, had been appointed satrap of the hilly districts on the Indus. 2 He must therefore have ordered his fleet to wait for him at some point on the river not far from the capital of Oxykanus. This point I would fix somewhere about Marija Dand, on the old Nara, below Kator and Tajal, as Mahorta, which I have identified with the chief city of Oxykanus, is about equidistant from Alor and Kator. Thence, descending the stream, he came on the fourth day to a town, through which there was a road to the kingdom of

1 Curtius, Vita Alex., ix. 8. " Alexander. . . . rursus amnem, in quo classem exspectare se jusserat, repetit. Quarto deinde die, secundo amne, pervenit ad oppidum, qua. iter in regnum erat Sabi."

2 Arrian, 'Anabasis,' vi. 16: <greek> ?

[p. 268]: Sambus. From Marija Dand, the point where I suppose that Alexander rejoined his fleet, the distance to the ruined city of Brahmana, or Brahmanabad, is 60 miles in a direct line by land, or 90 miles by water. As this distance could have been accomplished with ease in four days, I conclude that Brahmana was the actual city of Brahmans which is described by Alexander's historians. The king of this city had previously submitted, but the citizens withheld their allegiance, and shut their gates. By a stratagem they were induced to come out, and a conflict ensued, in which Ptolemy was seriously wounded in the shoulder by a poisoned sword. 1 The mention of Ptolemy's wound enables us to identify this city with that of Harmatelia, which Diodorus describes as the " last town of the Brahmans on the river." 2 Now, Harmatelia' is only a softer pronunciation of Brahma-thala, or Brahnana-sthala, just as Hermes, the phallic god of the Greeks, is the same as Brahma, the original phallic god of the Indians. But Brahmana was the old Hindu name of the city which the Muhammadans called Brahmanabad ; hence I conclude that the town of Brahmans captured by Alexander corresponds both in name and position with the great city of Brahmanabad.

The narrative of Arrian after the capitulation of Sindomana is unfortunately very brief. His words are, "he attacked and won a city which had revolted from him, and put to death as many of the Brahmans as fell into his hands, having charged them with being the authors of the rebellion."! This agrees with the

1 Curtius, Vita Alex., ix. 8.

2 Hist. Univers., xjii. 56.

3 ' Anabasis,' vi. 16. <greek>

[p. 269]:

statement of Diodorus, who mentions Alexander "was satisfied with punishing those who advised the resistance, and pardoned all the others." From a comparison of the three narratives, I infer that Harmatetia, or Brahmana, was in the dominions of Musikanus ; for Curtius states that the king of this city had previously submitted to Alexander, while Arrian says that he had revolted, and Diodorus adds1 that Alexander punished the advisers of the rebellion.

Now, all these facts apply to Musikanus, who had at first submitted, and then revolted, and was at last crucified, " and with him as many of the Brahmans as had instigated him to revolt." This identification is of some importance, as it shows that the dominions of Musikanus must have embraced the whole of the valley of the Indus down to the head of the Delta, with the exception of the two outlying districts of Oxykanus and Sambus, under the western mountains. This extension of his dominions explains the report which Alexander had previously received from the people, that the kingdom of Musikanus " was the richest and most populous throughout all India." It also explains how Sambus was at enmity with Musikanus, as the southern territories of the latter were bounded on the west by those of the former. The king of this city, where Ptolemy was wounded by a poisoned arrow, is called Ambiger by Justin, 2 which was probably the true name of Musikanus, the chief of the Musikani, in whose territory Brahmana was situated. It is much to be regretted that none of the names reserved by Ptolemy can be certainly identified with this city of the Brahmans. Parabali corresponds with

1 Hist. Univers., xvii. 56. 2 Justin, Hist., xii. 10.

[p. 270]: it in position, and partly also in name, as the first two syllables, Parab, are not very different from Baram, and the termination, ali, may represent thala of Brahmathala, or Harmatelia. After Ptolemy's time we know nothing of Brahmana until the Muhammadan conquest, a period of nearly six centuries. From the native histories, however, we learn that Brahmana was the chief city of one of the four governments1 into which Sindh was divided during the rule of the Rais dynasty, or from A.D. 507 to 642, and that it continued to be so until the accession of Dahir in A.D. 680, who made it the capital of the kingdom, after the destruction of Alor by the Indus. In A.D. 641 Sindh was visited by Hwen Thsang, whose account has already been noticed. He found the kingdom divided into the four districts, which for greater distinctness I have named Upper Sindh, Middle Sindh, Lower Sindh, and Kachh. The first has already been described in my account of Alor. The second, O-fan- cha, I have just identified with Brahmanabad. M. Stanislas Julien transcribes the Chinese syllables as Avanda, for which it is difficult to find an exact equivalent. But I have a strong suspicion that it is only a variation of the name of Brahmana, which was pro- nounced in many different ways, as Bahmana, Bahmana, Babhana, Babhana, Bambhana. 2 Speaking of Mansura, which we know was quite close to Brahmanabad, Ibn Haukal adds that the Sindhians call it Bamivan3 which Edrisi alters to Mirman. But in

1 Postans, Journ. Asiat. Soc. Bengal, 1838, pp. 93-96.

2 See Dhauli inscription of Asoka for Babhano, Babhana, and Bambhana : edicts iii. iv. viii. ix., in Journ. Royal Asiat. Soc, vol. xii.

3 Sir Henry Elliot, ' Muhammadan Historians of India,' p. 63.

4 Jaubert's ' Edrisi,' i. 162.

[p. 271]: his list of places in Sindh, Edrisi adds after Mansura the name of Wandan, or Kandan, 1 which I take to be only a various reading of Bamanwa, or, as the Sindhians would have pronounced it, Vamanwa, and Vanwa. The Chinese syllable fan, which is the wellknown transcript of Brahma, is a notable example of this very contraction, and tends to confirm the opinion that Avanda is but a slight variation of Bahmanwa, or Brahmanabad.

Shortly after the Muhammadan conquest Brahmana was supplanted by Mansura, which, according to Biladuri, was founded by Amru, the son of Muhammad bin Kasim, the conqueror of Sindh, 2 and named after the second Abasside Khalif Al Mansur, who reigned from A.D. 753 to 774. But according to Masudi 3 it was founded by Jamhur, the governor of Sindh, under the last Omnicad Khalif, a.d. 744 to 749, who named it after his own father Mansur. The new city was built so close to Brahmanabad that Ibn Haukal, Abu Rihan, and Edrisi, all describe it as the same place. Ibn Haukal' s words are, " Mansura, which in the Sind language is called Bamiwan." 4 Abu Rihan states that it was originally called Bamanhwa, and afterwards Hamanabad, for which we may read Bahmanabad, by simply adding an initial B, which must have been accidentally dropped. It was

1 Jaubert's 'Edrisi,' i. 160.

2 Reinaud, ' Fragments Arabes ; ' and Jaubert's ' Edrisi,' i. 162.

3 Sir Henry Elliot's 'Muhammadan Historians of India,' p. 57.

4 Sir Henry Elliot's ' Muhammadan Historians of India,' Dowson's edition, p. 34; and Jaubert's 'Edrisi,' i. 162. " Le nom de la ville (Mansura) est en Indien Miriman." In Gildemeister's 'Ibn Haukal,' this name is Tamirman, which is an obvious mistake for Bamiwan, or Bamanwas.

[p. 272]: situated on the eastern branch of the Mihran, or Indus, and was 1 mile in length, and the same in breadth, or just 4 miles in circuit. Its position is approximately fixed in the neighbourhood of Hala, by the number of days' journey in the routes to different places. It was 12 days from Multan, 8 from Kandabil, via Sehwan, and 6 days from Debal, via Manhabari, which was itself 4 days from Mansura. It was therefore at two-thirds of the distance from Multan to the mouth of the Indus, or very nearly in the same parallel as Hala.

Now in this very position the ruins of a large city have been discovered by Mr. Bellasis, to whose zeal and energy we are indebted for our knowledge of this interesting place. The ruins are situated near an old bed of the Indus, at 47 miles to the north-east of Haidarabad, 28 miles to the east or east-north-east of Hala, and 20 miles to the west of the eastern Nara. 1 The place is known as Bambhra-ka-thul, or " the Ruined Tower," from a broken brick tower which is the only building now standing. The present appearance of the site, as described by its discoverer, is "one vast mass of ruins, varying in size according to the size of the original houses." Its circumference, measured by a perambulator, is within a few yards of 4 miles. But besides the great mound of Bambhra-ka-thul, there is, at a distance of about 1-1/2 mile, "the distince and ruined city of Dolora, the residence of its last king, and 5 miles in another direction is the ruined city of Depur, the residence of his Prime Minister, and between these cities are the ruins of suburbs extending

1 Journ. Asiat. Soc, Bombay, v. 413; and Thomas's Prinsep, ii. 119. Eastwick's ' Handbook for Bombay,' p. 490.

[p. 273]: for miles far and wide into the open country." The great mound of Bambhraka-thul is " entirely surrounded with a rampart, mounted with numerous turrets and bastions." In the time of Akbar there were "considerable vestiges of this fortification," which Abul Fazl 1 says "had 140 bastions, one tanab distant from each other." The tanab was a measuring rope, which the emperor Akbar ordered to be changed for bambus joined by iron links. Its length was 60 Ilahi gaz, which, at 30 inches each, give 150 feet for the tanab; and this multiplied by 140, makes the circuit of the city 21,000 feet, or very nearly 4 miles. Now it will be remembered that Ibn Haukal describes Mansura as being 1 mile square, or 4 miles in circuit, and that Mr. Bellasis's measure of the circumference of the ruined mound of Bambhraka thul was within a few yards of 4 miles. From this absolute correspondence of size, coupled with the close agreement of position, which has already been pointed out, I conclude that the great mound of Bambhra ka thul represents the ruined city of Mansura, the capital of the Arab governors of Sindh. The Hindu city of Brahmana, or Brahmanabad, must therefore be looked for in the neighbouring mound of ruins now called Dilura, which is only 1.5 mile distant from the larger mound.

Mr. Bellasis, the discoverer of these ruins, has identified the great mound with Brahmanabad itself ;

1 ' Ayin Akbari,' ii. 115. Gladwyn's translation has 1400 bastions, which. would give to the city a circuit of 40 miles ; the MSS. have 149. The Ilahi gaz contained 41.5 Sikandari tanghas, and as the average breadth of 62 Sikandaris in my collection is 7234 inches, the length of the Ilahi gaz will be 30.0211 inches. Mr. Thomas, ii. 133, found exactly 30 inches.

[p. 274]: but to this it has been justly objected by Mr. Thomas1 that amongst the multitudes of mediaeval coins found during the excavations, " the number of Hindu pieces was very limited, and that even these seem to be casual contributions from other provinces, of no very marked uniformity or striking age." The local coins consist exclusively of specimens of the Arab governors of Sindh, with the name of Mansura in the margin ; and so far as I am aware, there is not a single piece that can be attributed to any of the Hindu rajas of Sindh. It is therefore to be regretted that Mr. Bellasis did not make more extensive excavations in the smaller mound of Dilura, which would probably have yielded some satisfactory evidence of its superior antiquity.

According to the native histories and traditions of the people, Brahmanabad was destroyed by an earth quake, in consequence of the wickedness of its ruler, named Dilu Rai. The date of this prince is doubtful. M'Murdo has assigned a.h. 140, or A.D. 757, 2 as the year in which Chota, the brother of Dilu, returned from his pilgrimage to Mekka ; but as Mausura was still a flourishing city in the beginning of the tenth century, when visited by Masudi and Ibn Haukal, it is clear that the earthquake cannot have happened earlier than A.D. 950. Dilu and Chota are said to have been the sons of Amir, the Rai or ruler of Brahmanabad. But it is difficult to believe that there were any Hindu chiefs in Brahmana during the rule of the Arab.-; in Mansura. The fact is that the same

1 Prinsep's Essays,' vol. ii. p. 121, where all the local coins are most carefully described and attributed,

2 Journ. Royal Asiat. Soc.. i. 28.

[p. 275]: stereotyped legend is told of all the old cities in the Panjab, as well as of those in Sindh. Shorkot, Harapa, and Atari, are all said to have been destroyed on account of the sins of their rulers, as well as Alor, Brahmana, and Bambhura. But the same story is also told of Tulamba, which we know to be false, as I have been able to trace its downfall to its desertion by the Ravi, at a very recent date. The excavations of Mr. Bellasis have shown conclusively that Brahmana was overwhelmed by an earthquake. The human bones " were chiefly found in doorways, as if the people were attempting to escape ; others in the corners of the rooms ; some upright, some recumbent, with their faces down, and some crouched in a sitting posture." 1 The city was certainly not destroyed by fire, as Mr. Richardson notes that he found no remains of charcoal or burnt wood, and that the old walls bore no traces of fire. On the contrary, he also found the human remains crushed in the corners of the rooms, as if the terror-stricken inhabitants, finding their houses falling about them, had crouched in the corners and been buried by the falling material. 2 Mr. Richardson also picked up a brick which had " entered cornerways into a skull, and which, when taken out, had a portion of the bone adhering to it." His conclusion is the same as that of Mr. Bellasis, " that the city was destroyed by some terrible convulsion of nature."

The local coins found in the ruins of Bambhra ka- tul belong to the Arab governors of Mansura, from the time of Jamhur, son of Mansur, the reputed founder of the city, down to Umar, the contemporary of

1 Journ. Asiat. Soc. Bombay, v. 417.

2 Ibid., V. 423.

[p. 276]: Masudi 1 It was therefore in existence during the whole of that time, or from A.D. 750 to 940, or even later. This agrees exactly with what I have already noted, that the city was still flourishing when visited by Masudi and Ibn Haukal in the first half of the tenth century ; and I would therefore assign its destruction to the latter half of that century, and not earlier than A.D. 970. It is true that Mansura is mentioned by Abu Rihan in the beginning of the next century, and at a still later period by Edrisi, Kazvini, and Rashid-ud-din ; but the last three were mere compilers, and their statements accordingly belong to an earlier age. Abu Rihan, however, is entirely original, and as his knowledge of the Indian language gave him special facilities for obtaining accurate information, his evidence is sufficient to prove that Mansura was still existing in his time. In speaking of the itinerary of Sindh, he says, 2 " From Aror to Bahmanwa, also named el Mansura, is reckoned 20 parasangs; from thence to Loharani, at the mouth of the river, 30 parasangs." Mansura therefore still existed when Abu Rihan wrote his work, about A.D. 1031 ; but as it is mentioned by only one author in the campaigns of Mahmud of Ghazni, it is almost certain that it no longer existed as a great fortress, the capital of the country, otherwise its wealth would have attracted the cupidity of that rapacious conqueror. I conclude, therefore, that Mansura was already very much decayed before the accession of Mahmud, and that the earthquake -which levelled its walls and overthrew its houses, must have happened some time before

1 Thomas in Prinsep's ' Essays,' ii. 113. 2 Reinaud, ' Fragments Arabes,' etc. p. 113.

[p. 277]: the beginning of the eleventh century. It is probable that most of the inhabitants who escaped the great catastrophe would have returned to the ruined city to look after their buried property, and that many of them again reared their houses on the old sites. But the walls of the city were fallen, and there was no security ; the river was gradually failing, and there was a scarcity of water ; and the place was altogether so much decayed, that even in a.h. 416, or A.D. 1025, when the conqueror of Somnath returned through Sindh, the plunder of Mansura was not sufficient to tempt him out of his direct march ; so he passed on by Sehwan to Ghazni, leaving the old capital unvisited, and even unnoticed, unless we accept the solitary statement of Ibn Athir, that Mahmud on this occasion appointed a Muhammadan governor to Mansura.

3. Lower Sindh or Lar

[p. 277]:The district of Pitasila, or Lower Sindh, is described by Hwen Thsang as being 3000 li, or 500 miles, in circuit, which agrees almost exactly with the dimensions of the Delta of the Indus from Haidarabad to the sea, including a small tract of country on both sides, extending towards the desert of Umarkot on the east, and to the mountains of Cape Monz on the west. Within these limits the dimensions of Lower Sindh are as follows. From the western mountains to the eighbourhood of Umarkot, 160 miles; from the same point to Cape Monz, 85 miles ; from Cape Monz to the Kori mouth of the Indus, 135 miles; and from the Kori mouth to Umarkot, 140 miles ; or altogether 520 miles. The soil, which is described as sandy and

[p. 278]: salt, produced plenty of corn and vegetables, but very few fruits and flowers, which is true of the Delta to the present day.

In the time of Alexander, the only place of note in the Delta was Patala ; but he is said to have founded several towns himself1 during his long stay in Lower Sindh, waiting for the Etesian winds to start his fleet. Unfortunately the historians have omitted to give the names of these places. Justin alone notes that on his return up the Indus he built the city of Barce, 2 to which I shall hereafter refer. Ptolemy has preserved the names of several places, as Barbara, Sousikana, Bonis, and Kolaka, of which the first is most probably the same as the Barbarike emporium of the 'Periplus,' and perhaps also the same as the Barce of Justin. In the time of the author of the 'Periplus,' the capital of Lower Sindh was Minnagara, which the foreign merchants reached by ascending the river from Barbarike. In the middle of the seventh century, Hwen Thsang mentions only Pitasila, or Patala. But in the beginning of the eighth century, the historians of Muhammad bin Kasim's expedition add the names of Debal and Nirankot to our scanty list, which is still further increased by the Arab geographers of the tenth century, who place Manhatara, or Manhabari, or Manjabari, 3 to the west of the Indus, and two days' journey from Debal, at the point where the road from Debal crosses the river. The position of these places I will now investigate in their order from north to

1 Curtius, Vita Alex., ix. 10: " Interim et urbes plcrasque ondidit."

2 Hist., xii. 10.

3 Sir Henry Elliot, ' Muhammadan Historians of India,' Dowson's edition, i. 35, quoting Ibn Haukal.

[p. 279]: south, beginning with Patala, at the head of the Delta.

Patala or Nirankot

The position of Nirankot is fixed at Haidarabad by the concurrent testimony of M'Murdo, Masson, Burton, and Eastwiek. 1 Sir Henry Elliot alone places it at Jarak, as he thinks that that locality agrees better with the descriptions of the native historians. But as Haidarabad is the modern name of the city, which the people still know as Nirankot, there would seem to be no doubt of its identity with the Nirun, or Nirunkot, of the Arab historians and geographers. Its position is described by Abulfeda as 25 parasangs from Debal, and 15 parasangs from Mansura, which accords with the less definite statements of Istakhri and Ibn Haukal, who simply say that it was between Debal and Mansura, but nearer to the latter. It was situated on the western bank of the river, and is described as a well-fortified but small town, with few trees. Now, Haidarabad is 47 miles from the ruined city of Brahmanabad, or Mansura, and 85 miles from Lari-bandar, which I will presently show to have been the most probable position of the ancient Debal ; while Jarak is 74 miles from Brahmanabad, and only 60 miles from Lari-bandar. The position of Haidarabad, therefore, corresponds much better with the recorded distances than that of Jarak. At present the main channel of the Indus runs to the west of Haidarabad, but we know that the Phuleli, or eastern branch,

1 M'Murdo in Journ. Royal Asiat. Soc., i. 30; Masson, 'Travels,' i. 463 ; Burton, ' Sindh,' pp. 131, 376 ; and Eastwiek, ' Handbook for Bombay,' p. 483. See Map No. IX.

[p. 280]: was formerly the principal stream. According to M'Murdo, 1 tlic change of the main stream to the westward of Haidarabad took place prior to a.h. 1000, or A.D. 1592, and was coincident with the decay of Nasirpur, which was only founded in a.h. 751, or A.D. 1350. As Nasirpur is mentioned by Abul Fazl 2 as the head of one of the subdivisions of the province of Thatha, the main channel of the Indus must have flowed to the eastward of Nirunkot or Haidarabad at as late a date as the beginning of the reign of Akbar.

Nirunkot was situated on a hill, and there was a lake in its neighbourhood of sufficient size to receive the fleet of Muhammad Kasim. Sir Henry Elliot identifies the former with the hill of Jarak, to the west of the Indus, and the latter with the Kinjur lake, near Helai, to the south of Jarak. But the Kinjur lake has no communication with the Indus, and therefore could not have been used for the reception of the fleet, which at once disposes of the only special advantage that Jarak was supposed to possess over Haidarabad as the representative of Nirunkot. Sir Henry 3 admits "that the establishment of its locality depends chiefly upon the sites which are assigned to other disputed cities, more especially to Debal and Mansura." The former he identifies with Karachi, and the latter with Haidarabad; and con- sistently with these emplacements he is obliged to fix Nirunkot at Jarak. But since he wrote his ' Appendix to the Arabs in Sindh,' the ancient city of

1 Journ. Royal Asiat. Soc, i. 236.

2 ' Ayin Akbari,' ii. 272.

3 Sir H. Elliot's ' Muhammadan Historians of India,' Dowson's edition, i. 400.

[p. 281]: Bambhra-ka-Thul lias been found by Mr. Bellasis in the very position that was long ago pointed out by M'Murdo as the site of Brahmanabad. Its identification as the site of the famous cities of Mansura and Brahmanabad leaves Haidarabad, or the ancient Nirankot, available as the true representative of the Nirunkot of Biladuri and the Chach-nama. Its distance of 47 miles from Bambhra-ka-tul, and of 85 miles from Lari-bandar, agree almost exactly with the 15 and 25 parasangs of Abulfeda. It is also situated on a hill, so that it corresponds in position, as well as in name, with Nirunkot. The hill, called Ganja, is 1.5 mile long, and 700 yards broad, with a height of 80 feet. 1 The present fort was built by Mir Ghulam Shah in a.h. 1182, or a.d. 1768. 2 About one-third of the hill, at the southern end, is occupied by the fort, the middle portion by the main street and straggling houses of the city, and the northern end by tombs.

In A.D. 641, when the Chinese pilgrim Hwen Thsang visited Sindh, he travelled from Koteswara, the capital of Kachh, a distance of 700 li, or 117 miles, due north to Pi-to-shi-lo, 3 from whence he proceeded 300 li, or 50 miles, to the north-east, to, which I have already identified with Brahmanabad. M. Julien renders the Chinese syllables by Pitasila, but I should prefer Patasila, or the "flat rock," which is an accurate description of the long flat-topped hill on which Haidarabad is situated.

This name recalls that of Patalpur, which, according

1 Wood, 'Journey to the Source of the Oxus,' p. 30. 2 M'Murdo, Journ. Royal Asiat. Soc, i. 234. 3 ' [[Hiouen Thsang[[,' iii. 180.

[p. 282]:

to Burton, 1 was an old appellation of Haidarabad, or Nirankot ; and as this city is exactly 120 miles to the north of Kotesar, in Kachh, and 47 miles to the south- west of Brahmanabad, I have no hesitation in identifying it with the Pitasila of the Chinese pilgrim. The size of the hill also, which is 1.5 mile in length, by 700 yards in breadth, or upwards of 3 miles in circumference, corresponds very closely with the dimensions of Pitasila, which, according to Hwen Thsang, was 20 li, or 3-1/3 miles, in circuit.

The names of Patalpur and Patasila further suggest the probability that Haidarabad may be the Pattala of Alexander's historians, which they are unanimous in placing near the head of the Delta. Now, the present head of the Delta is at the old town of Mattari, 12 miles above Haidarabad, where the Phuleli separates from the main channel of the Indus. But in ancient times, when the main stream, which is now called Purana, or the "Old River," flowed past Alor and Brahmanabad to Nirunkot, the first point of separation of its waters was either at Haidarabad itself, past which a branch is said to have flowed by Miani to Trikal, or 15 miles to the south-east of it where the Phuleli now throws off the Guni branch to the south, and then proceeds westerly to join the present stream of the Indus at Trikal. The true head of the old Delta was therefore either at Haidarabad itself, or 15 miles to the south-east of it, where the Guni, or eastern branch of the Indus, separated from the Phuleli, or western branch.

Now, the position of Patala can be determined by several independent data: —

1 ' Sindh,' chap, i, note 7.

[p. 283]: 1st. According to Ptolemy, the head of the Delta was exactly midway between Oskana and the eastern mouth of the Indus, called Lonibare ostium. This fixes Patala at Haidartlbad, which is equidistant from the capital of Oxykanus., that is, from Mahorta near Larkana, and the Kori., or eastern mouth of the Indus, which is also the mouth of the Loni river, or Lonibare ostium.

2nd. The base of the Delta was reckoned by Aristobulus at 1000 stadia, or 115 miles; by Nearchus at 1800 stadia, and by Onesikritus at 2000 stadia1 But as the actual coast line, from the Ghara mouth on the west, to the Kori mouth on the east, is not more than 125 miles, we may adopt the estimate of Aristobulus in preference to the larger numbers of the other authorities. And as Onesikritus states that all three sides of the Delta were of the same length, the distance of Patala from the sea may be taken at from 1000 stadia, or 115 miles, up to 125 miles. Now, the distance of Haidarabad from the Ghara, or western mouth of the Indus, is 110 miles, and from the Kori, or eastern mouth, 135 miles, both of which agree sufficiently near to the base measurement to warrant the descriptions of Onesikritus that the Delta formed an equilateral triangle. Consequently, the city of Patala, which was either at or near the head of the Delta, may be almost certainly identified with the present Haidarabad.

3rd. From a comparison of the narratives of Arrian and Curtius, it appears that the Raja of Patala, having made his submission to Alexander at Brahmana, or the city of Brahmans, the conqueror sailed leisurely

1 Strabo, Geogr., xv. i. 33.

[p. 284]: down tile river for three days, when he heard that the Indian prince had suddenly abandoned his country and fled to the desert. 1 Alexander at once pushed on to Patala. Now, the distance from Brahmanabad to Haidarabad is only 47 miles by the direct land route ; but as the old bed of the Indus makes a wide sweep round by Nasirpur, the route along the river bank, which was doubtless followed by the army, is not less than 55 miles, while the distance by water must be fully 80 miles. His progress during the first three days, estimated at the usual rate of 10 or 12 miles by land, and 18 or 20 miles by water, would have brought him within 19 miles of Haidarabad by land, and 26 miles by water, which distance he would have easily accomplished on the fourth day by a forced march. From Patala he proceeded down the western branch of the river for a distance of 400 stadia, or 46 miles, when his naval commanders first perceived the sea breeze. This point I believe to have been Jarak, which is 30 miles below Haidarabad by land, and 45 miles, or nearly 400 stadia, by water. There Alexander procured guides, and, pressing on with still greater eagerness, on the third day he became aware of his vicinity to the sea by meeting the tide. 2 As the tides in the Indus are not felt more than 60 miles from the sea, I conclude that Alexander must then have reached as far as Bambhra, on the Ghara, or western branch of the river, which is only 35 miles from the sea by land, and about 50 miles by water. Its distance from Jarak by land is 50 miles, and by

1 Arrian, ' Anabasis,' vi. 17 ; Curtius, Vita Alex., ix. 8, 28, says that lie fled to the mountains.

2 Curtius, Vita Alex., ix. 9, 29.

[p. 285]: water 75 miles, which the fleet might have easily accomplished by the third day. From these details it is clear that Patala must have been at a considerable distance from the sea, that is, not less than the length of the tidal reach, plus three days' sail on the river, plus 400 stadia. These distances by land are respectively 33 miles, 50 miles, and 30 miles, or altogether 113 miles, which corresponds almost exactly with the measurement of Aristobulus of 1000 stadia, or 115 miles.

As these three independent investigations all point to the same place as the most probable representative of Patala, and as that place is called Patasila by Hwen Thsang in the seventh century, and is still known as Patalpur, I think that we have very strong grounds for identifying Haidarabad with the ancient Patala.

In his account of the Indus, Arrian 1 says, "this river also forms a delta by its two mouths, no way inferior to that of Egypt, which, in the Indian language, is called Pattala." As this statement is given on the authority of Nearchus, who had ample opportunities during his long detention in Sindh of intercourse with the people, we may accept it as the general belief of the Sindhians at that time. I would therefore suggest that the name may have been derived from Patala, the "trumpet-flower" (Bignonia suaveolens), in allusion to the " trumpet " shape of the province included between the eastern and western branches of the mouth of the Indus, as the two branches, as they approach the sea, curve outwards like the mouth of a trumpet.

I cannot close the discussion on the site of this

1 ' Indica,' p. 2.

[p. 286]: ancient city without noticing another name of which the conflicting accounts appear to me to have a confused reference to Nirunkot. This name is the Piruz of Istakhri, the Kannazhur of Ibn Haukal, and the Firahuz of Edrisi. According to Istakhri, Piruz was 4 days' journey from Debal, and 2 days from Mehabari, which was itself on the western bank of the Indus, at 2 days' journey from Debal. Ibn Haukal and Edrisi agree that the road to Kannazbur, or Firabuz, lay through Manhabari, or Manjabari, which was on the western bank of the Indus, at 2 days from Debal ; but they make the whole distance beyond Debal 14 days instead of 4. Now, Ibn Haukal and Edrisi place their city in Mekran, a position which they were almost forced to adopt by their long distance of 14 days, although the first two days' journey lie exactly in the opposite direction from Mekran. But if we take the shorter distance of 4 days from Debal, which is found in Istakhri, the earliest of the three geographers, the position of their unknown city will then accord exactly with that of Nirankot. Debal I will hereafter identify with an old city near Laribandar and Manhabari with Thatha, which is just midway between Lari-bandar and Haidarabad. Now, Ibn Haukal specially notes that Manjabari was situated "to the west of the Mihran, and there any one who proceeds from Debal to Mansura will have to pass the river, the latter place being opposite to Manjabari." 1 This extract shows that Manjabari was on the western branch of the Indus, and therefore on the high-road to Nirankot as well as to Piruz, or Kannezbur, or Firabuz. I would therefore suggest that the first of

1 Prof. Dowson's edition of Sir H. Elliot's Hist, of India, i. 37.

[p. 287]: these names, which is thus mentioned in conjunction with Manhabari might possibly be intended for Nirun, and the other two for Nirunkot, as the alterations in the original Arabic characters required for these two readings are very slight. But there was certainly a place of somewhat similar name in Mekran, as Biladuri records that Kizbun in Mekran submitted to Muhammad Kasim on his march against Debal. Comparing this name with Ibn Haukal's Kannazbur, 1 and Edrisi's Firabuz, I think it probable that they may be intended for Panjgur, as suggested by M. Reinaud. The 14 days' journey would agree very well with the position of this place.


The little town of Jarak is situated on an eminence overhanging the western bank of the Indus, about midway between Haidarabad and Thatha. Jarak is the present boundary between Vichalo, or Middle Sindh, and Lar, or Lower Sindh, which latter I have been obliged to extend to Haidarabad, so as to include the Patala of the Greeks and the Pitasila of the Chinese pilgrim, within the limits of the ancient Delta. This is perhaps the same place as Khor, or Alkhor, a small but populous town, which Edrisi places between Manhabari and Firabuz, that is, between Thatha and Nirunkot. Three miles below Jarak there is another low hill covered with ruins,

1 Prof. Dowson's edition of Sir Henry Elliot's Hist. of India, i. 40. Ibn Haukal : Kannazhur. At page 29 lie gives Istakhri's name as Kannazbun, which Mordtmann reads Firiun. The most probable explanation of these differences is some confusion in the Arabic characters between the name of Nirun and that of the capital of Mekran.

[p. 288]: which the people call Kafir kot, and attribute to Raja Manjhira 1 The principal ruin is a square basement ornamented with fiat pilasters at regular distances. This is supposed to be the remains of a temple. Amongst the ruins were found some fragments of Buddhist statues ; and, at a short distance from the hill, an inscription in early Indian characters, of which I can read only the words putrasa and Bhagavatasa, and a few letters in different parts ; but these are sufficient to show that the inscription is Buddhist, as well as the other remains.

Minnagar, Manhabari, or Thatha

The city of Thatha is situated in a low swampy valley, 3 miles from the western bank of the Indus, and 4 miles above the separation of the Bagar, or western branch, from the Sata, or main stream of the river. Littlewood remarks that " the mounds of rubbish upon which the houses are piled slightly raise its site above the level of the valley." 2 The place was visited by Captain Hamilton in A.D. 1699, who describes it 3 as situated on a spacious plain about 2 miles from the Indus. It is highly probable, therefore, that the town originally stood on the bank of the river, which has been gradually receding from it. Its name also would seem to point to the same conclusion, as thattha means a " shore or bank," so that Nagar-Thatha, which is the common name of the place, would mean the " city on the river bank." Its date is not certainly known ; but M'Murdo, who is gene-

1 ' Bombay Journal,' v. 356.

2 Journey to the Source of the Oxus,' p. 17.

3 'New Account of the East Indies,' i. 123.

[p. 289]: rally very accurate, states that it was founded in the year A.H. 900, or A.D. 1495, by Nizam-ud-din Nanda, the Jam, or ruler of Sindh.

Before his time, the chief city of Lower Sindh was Saminagar, the capital of the Samma tribe, which stood on a rising ground, 3 miles to the north-west of the site of Thatha. M'Murdo refers its foundation to the time of Ala-ud-din of Delhi, who reigned from A.H. 695-715, or A.D. 1295 to 1315.

Of a still earlier date is the great fort of Kalyan-kot, or Tughlakabad, which stands on the limestone hill, 4 miles to the south-west of Thatha. Its second name was derived from Ghazi Beg Tughlak, who was the governor of Multan and Sindh, during the latter part of Ala-ud-din' s reign, in the beginning of the fourteenth century.

The site of Thatha itself is admitted to be modern, but those of Saminagar and Kalyan-kot are said to be of great antiquity. This belief of the people is no doubt true, as the position at the head of the inferior Delta commanded the whole traffic of the river, while the hill-fort gave security. Lieut. Wood remarks1 that the site of Thatha is so advantageous for commercial purposes that it is probable that a mart has existed in its neighbourhood from the earliest times. " But," he judiciously adds, " as the apex of the Delta is not a fixed point, the site of this city must have varied as the river changed." This change of site would naturally have entailed a change of names ; and I am therefore led to believe that Thatha was the actual position of the Manhabari of the Arab geographers, and of the Minnagara of the author of the ' Periplus.'2

1 'Oxus,' p. 20. 2 See Map No. IX.

[p. 290]:

Manhabari is described by all the authorities as situated on the western bank of the Indus, at 2 days' journey from Debal. Now, this is the very position of Thatha, which is on the western bank of the Indus, at 40 miles, or 2 days' journey, from Lari-bandar, which, as I will presently show, was almost certainly within a few miles of the famous city of Debal. The name of Manhabari is variously written as Mehabari, and Manjabari, for which I would suggest that we might perhaps read Mandabari, or Mandawari, the " city of the Mand” tribe, just as Saminagar was the "city of the Samma," tribe. This derivation of the name is supported by the fact that the Mand tribe have occupied Lower Sindh in great numbers from the beginning of the Christian era. Edrisi 1 describes the Mand as a numerous and brave tribe, who occupied the desert on the borders of Sindh and India, and extended their wanderings as far as Alor on the north, Mekran on the west, and Mamehel (or Umarkot) on the east. Ibn Haukal 2 records that "the Mands dwell on the banks of the Mihran, from the boundary of Multan to the sea, and in the desert between Mekran and Famhal (or Umarkot). They have many cattle-sheds and pasturages, and form a large population." Rashid-ud-din 3 locates them in Sindh at a still earlier period. According to his account, Med and Zat, two descendants of Ham, the son of Noah, were the progenitors of the people of Sindh prior to

1 Geogr., i. 163.

2 In Elliot, ' Muhammadan Historians of India,' i. 67 ; and in Gildemeister, ' De Rebus Indicis,' p. 172, where he gives Kamuhal as the eastern limit of their wanderings.

3 Reinaud, ' Fragments Arabes,' etc., p. 25.

[p. 291]: the Mahabharata. The name is variously written as Mer, Med, Mand, in all of which forms it is found even at the present day. To these I would add Mind, which is the form of the name given by Masudi. 1 I have already identified this people with the Medi and Mandrueni of the classical writers ; and as their name is found in northern India from the beginning of the Christian era downwards, and not before that time, I conclude that the Mandrueni and Iatii of the Oxus, who are coupled together by Pliny, must be the Sacae Indo-Scythians, who occupied the Panjab and Sindh, and who under the name of Mands and Zats of the early Muhammadan authors, were in full possession of the valley of the Indus towards the end of the seventh century.

To show that the various spellings of the name are but natural modes of pronunciation, I can refer to the two large maps of the Shahpur and Jhelam, districts, which have been published within the last few years by the Surveyor-General of India. In the latter the name of a village on the Jhelam, 6 miles above Jalalpur, is spelt Meriala, and in the former Mandiali. Abul Fazl calls the same place Merali, while Ferishta names it Meriala. Lastly, Wilford's surveyor, Mogal Beg, writes Mandyala, which is also the form that I received from two different persons, while in General Court's map it is spelt Mamriala.

To this people I refer the name of Minnagar, or " city of the Min," which was the capital of Lower Sindh in the second century of the Christian era. That Min was a Scythian name we know from its

1 Sir H. M. Elliot, ' Muhammadan Historians of India,' Dowson's edition, i. 57.

[p. 292]: occurrence in the list of Isidor of Xharax as one of the cities of Sakastene, or Sejistan. The appearance of the name in Sindh would alone be sufficient to suggest the presence of Scythians ; but its connection with them is placed beyond all doubt by the mention that the rulers of Minnagara were rival Parthians, who were mutually expelling each other. 1 These Parthians were Dahae Scythians from the Oxus, who gave the name of Indo-Scythia to the valley of the Indus, and whose mutual rivalry points to their identity with the rival Meds and Jats of the Muhammadan authors.

The actual position of Minnagar is unknown, and we have but few data to guide us in attempting to fix its site. As it is not found in Ptolemy, who wrote in the first half of the second century, I infer either that the new name had not then been imposed on the capital, or what is more probable, that Ptolemy has inserted only the old name. If I am right in identifying Min-nagara, or the "city of the Min," with Mand-abari, or the "place of the Mand, there can be little doubt that the great Indo-Scythian capital was at Thatha. Edrisi 2 describes Manhabar as situated on a low plain, and surrounded with gardens and running water. Captain Hamilton3 gives the same description of Thatha, which, he says, "stands in a spacious plain, and they have canals cut from the river, that bring water to the city, and some for the use of their gardens." According to the author4 of

1 Peripl. Mar. Eryth. ; in Hudson's Geogr. Vet., i. 22.

2 Geogr., i. 164.

3 ' New Account of the East Indies,' i. 123.

4 Hudson, Geogr. Vet., i. 22.

[p. 293]: the ' Periplus,' the merchant vessels anchored at the emporium of Barbarike, where the goods were un-loaded, and conveyed to the capital by the river. Just so in modern times the ships anchored at Lari- Bandar, while the merchants carried their goods to Thatha either by land or by water. The position of Minnagar is too vaguely described as " inland," 1 to be of any use in its determination. If it was, as I suppose, at Thatha, then it may perhaps be identified with Ptolemy's Sousikana, which I would interpret as Susi-gama, or the "town of the Su, tribe," an etymology which is supported by the fact that the Mands, or Meds, were a branch of the great horde of Sus, or Abars, who gave one name to Susiana, at the mouth of the Euphrates, and the other to Abiria, at the mouth of the Indus. I should mention, however, that according to M'Murdo, 2 " Minagar was one of the cities dependent on Multan in the twelfth century, and was the possession of a chief by caste an Agri, and descended from Alexander. It was situated on the Lohana Darya, not far from Bahmana, in the parganah now called Shehdadpur.” It is a suspicious ci cumstance that this passage has not been verified either by Postans or by Elliot. The latter, who constantly refers to his own MS. of the ' Tohfat-ul-Kiram,' quotes; 3 this notice of Minagar at second-hand from M'Murdo. I may add that the Agari is a well-known caste, of low degree, who are employed in the manu-

1 The words are, <greek>, which can only mean " inland and beyond" Barbarike.

2 Journ. Royal Asiat. Soc, i. 31 ; and again at p. 233, quoting the Tohfat-ul Giram.

3 ' Muhammadan Historians of India,' Dowson's edition, i. 66.

[p. 294]: facture of salt. I am therefore not inclined to admit that this petty place could have any connection with the great capital of Indo-Scythia. On the contrary, I am disposed to look upon this name of Min-nagara as meaning simply the city of Min.

Barbarike-Emporiam, or Bhambura

[p. 294]: The ruined town of Bambhora, or Bhambura, is situated at the head of the Ghara creek, which is "supposed by the natives to be the site of the most ancient seaport in Sindh." 1 "Nothing now remains but the foundations of houses, bastions, and walls," but about the tenth century Bhambhura was the capital of a chief named Bhambo Raja. According to the traditions of the people, the most westerly branch of the Indus once flowed past Bhambura. It is said to have separated from the main river just above Thatha, and M'Murdo2 quotes the ' Tabakat-i-Akbari ' for the fact that in the reign of Akbar it ran to the westward of Thatha. To the same effect Sir Henry Elliot; 3 quotes Mr. N. Crow, who was for many years the British Resident at Thatha. Writing in A.D. 1800, Crow says, "By a strange turn that the river has taken within these five-and-twenty years just above Tatta, that city is flung out of the angle of the inferior Delta, in which it formerly stood, on the main land towards the hills of Biluchistan." From these statements it would appear that the Ghara river was the most westerly branch of the Indus down to the latter half of the last century. But long before that time,

1 Eastwick, ' Handbook of Bombay,' p. 481.

2 Journ. Royal Asiat. Soc, i. 25. See Map No. IX.

3 Muhammadan Historians of India, Dowson's edition, i. 399.

[p. 295]: according to M'Murdo, it had ceased to be a navigable stream, as both. Bhambur and Debal were deserted about A.D. 1250, on account of the failure of the river. 1 My own inquiries give the same date, as Debal was still occupied when Jalaladdin of Khwarazm invaded Sindh in A.D. 1221, 2 and was in ruins in A.D. 1333, when Ibn Batuta visited Lahari Bandar, which had succeeded Debal as the great port of the Indus.

M'Murdo quotes native authors to show that this western branch of the Indus was called the Sagara river, which, he thinks, may .be identified with the Sagapa Ostium of Ptolemy, which was also the most westerly branch of the Indus in his time. It is therefore quite possible, as supposed by M'Murdo, that this was the very branch of the Indus that was navigated by Alexander. From the latest maps, however, it appears that about midway between Thatha and Ghara this channel threw off a large branch on its left, which flowed parallel to the other for about 20 miles, when it turned to the south and joined the main channel just below Lari-bandar. Now this channel passes about 2 or 3 miles to the south of Bhambura, so that the town was also accessible from the Piti, the Phundi, the Kyar; and the Pintiani mouths of the river. I am therefore inclined to identify Bhambura not only with the town of Barke, which Alexander built on his return up the river, as stated by Justin, but also with the Barbari of Ptolemy, and the Bar- barike Emporium of the author of the ' Periplus.' The last authority describes the middle branch of the

1 Jonrn. Royal Asiat. Soc, i. 25 and 232.

2 Rashid-ud-din in Elliot, Dowson's edition, i. 26.

[p. 296]: Indus as the only navigable channel in his time up to BarbariJce* all the other six channels being narrow and full of shoals. This statement shows that the Ghara river had already begun to fail before a.d. 200. The middle mouth of the river, which was then the only navigable entrance, is called Khariphon Ostium by Ptolemy. This name I would identify with the Kydr river of the present day, which leads right up to the point where the southern branch of the Ghara joins the main river near Lari-bandar.

From this discussion I conclude that the northern channel of the Ghara was the western branch of the Indus, which was navigated by Alexander and Nearchus ; and that before A.D. 200, its waters found another channel more to the south, in the southern Ghara, which joins the main stream of the Indus just below Lari-bandar. By this channel, in the time of the author of the ' Periplus,' the merchant vessels navigated the Indus up to Barbarike, where the goods were unloaded, and conveyed in boats to Minnagar, the capital of the country. But after some time this channel also failed, and in the beginning of the eighth century, when the Arabs invaded Sindh, Debal had become the chief port of the Indus, and altogether supplanted Bhambura, or the ancient Barbarike. But though the Ghara river was no longer a navigable channel, its waters still continued to flow past the old town down to the thirteenth century, about which time it would appear to have been finally deserted.

1 Hudson, Geogr. Vet., i. 22.

[p. 297]:

Debal Sindhi or Debal

[p. 297]: The position of the celebrated port of Debal, the emporium of the Indus during the middle ages, is still unsettled. By Abul Fazl and the later Muhammadan writers, Debal has been confounded with Thatha; but as Debal was no longer in existence when they wrote, I conclude that they were misled by the name of Debal Thatha, which is frequently applied to Thatha itself. Similarly, Brahmana, or Brahmanabad, was called Debal Kangra, and the famous seaport of Debal was named Debal Sindhi. But Diwal, or Debal, means simply a temple, and therefore Debal Sindhi means the temple at, or near, the town of Sindhi. Major Burton says that the shawls of Thatha are still called Shal-i-Debali, but this only proves that Debal was the place where the merchants procured the Thatha shawls. Just so the name of Multani-matti, that is Multan clay, or Armenian bole, is derived from the place where the merchants obtain the article, as the clay is actually found in the hills to the west of the Indus, beyond Dera Ghazi Khan. So also Indian-ink is named from India, where the merchants first obtained it, although, as is now well known, it is all manufactured in China. Sir Henry Elliot, who is the last inquirer into the geography of Sindh, places Debal at Karachi ; but admits that Lari-bandar "is the next most probable site after Karachi." 1 But I incline to the opinion of Mr. Crow, who was for many years the British resident in Sindh, that Debal occupied a site between Karachi and Thatha. His opinion is entitled to special weight, as he is

1 ' Sindh,' pp. 222 and 224.

[p. 298]:

admitted by M'Murdo and Elliot to have " combined much, discrimination with ample opportunities of local inquiry." Sir Henry quotes the Chach-nama for the fact that "the Serandip vessels were in their distress driven to the shore of Debal," to show that the port must have been close to the sea. There they were attacked by pirates of the Tangamara tribe, who occupied the seacoast from Karachi to Lari-bandar. This statement shows that if Debal cannot be identified either with Karachi or with Lari-bandar, it must be looked for somewhere between them.

In favour of Karachi Sir Henry quotes Biladuri, who records that in the year A.H. 15, or A.D. 636, Hakim dispatched his brother Mughira on an expedition to the Bay of Debal. But as the city of Lyons is not on the shore of the Gulf of Lyons, so it does not necessarily follow that Debal was on the shore of the Bay of Debal. In fact it is described by Ibn Khordadbeh as being 2 parsangs from the mouth of the Mihran, which is still further extended to 2 days' journey by Masudi. 1 But as Debal was situated on the Indus, it cannot be identified with Karachi, which is on the seacoast beyond the mouth of the river. All our authorities agree in stating that it was on the west side of the Mihran, 2 that is of the main stream of the river, or Baghar, which flows past Lari-bandar, and discharges itself into the sea by several different mouths named the Piti, the Phundi, the Kyar, and the Pintiani. But M'Murdo also quotes the native

1 Elliot, Muhammadan Historians of India, Dowson's edition, i. 53-57.

2 These will be found in Elliot's Muhammadan Historians of India, by Dowson, i. 61 ; ' Istakhri," i. 65 ; ' Ashkal-ul-Bilad,' i. 65, note Ibn Haulial. See also Gildemeister, ' De Rcbus Indicis,' p. 205, for Kazvini.

[p. 299]: authorities to show that it was on the Sagara branch of the Indus, which flowed past Bhambura. According to these accounts, Debal must have been situated on the western bank of the Baghar river, below the junction of the southern branch of the Ghara, or Sagara, branch. Its position may therefore be fixed approximately at the point of junction, which is 5 miles to the north of Lari-bandar, 17 miles to the south-west of Bhambura, and about 30 miles from the Piti and Pintiani mouths of the river. This position also fulfils the other condition quoted by Sir Henry Elliot, that Debal was between Karachi and Lari Bandar, in the territory of the Tangamara tribe of pirates. It further agrees with the position assigned to it by Mr. Crow, who places it between Karachi and Thatha, which is an exact description of the locality following the course of the river, which is the only course that can be taken, as Debal was situated amongst the intersecting streams of the Delta.

Unfortunately, this part of the Delta has not yet been minutely explored ; and to this cause I would attribute our ignorance of the remains of an ancient city, which were noticed by Ibn Batuta in A.D. 1333 in the very position which I have assigned to Debal. 1 As his statement is of great importance, I will quote the passage at full length: — "I then proceeded by the Sind to the city of Lahari, which is situated upon the, shores of the Indian Sea, where the Sind joins it. It has a large harbour, into which ships from Persia, Yemen, and other places put. At a few miles from this city are the ruins of another, in which stones in the shape of men and beasts almost innumerable are

1 ' Travels,' by Dr. Lee, p. 102.

[p. 300]: to be found. The people of this place think that it is the opinion of their historians that there was a city formerly in this place, the greater part of the inhabitants of which were so base that God transformed them, their beasts, their herbs, even to the very seeds, into stones ; and indeed stones in the shape of seeds are here almost innumerable." This large ruined city, with its stones in the shape of men and beasts, I take to be the remains of the once great emporium of Debal. According to M'Murdo, the people of Debal moved to Lari-bandar, 1 and according to Captain Hamilton, Lari-bandar possessed "a large stone fort," for the protection of merchants against the Biluchis and Makranis. It is, I think, a very fair and legitimate deduction that the people who deserted Debal removed the materials of their old city for the construction of the new one, and therefore that the stones of the fort of Lari-bandar were brought from the deserted city of Debal, the remains of which excited the curiosity of Ibn Batuta in A.D. 1333.

This statement of Ibn Batuta I would connect with the curious account of an Indian city in the ' Arabian Nights,' which is found in the story of Zobeide. According to the common edition, this lady sailed from the port of Bassora, and after twenty days anchored in the harbour of a large city in India, where on landing she found that the king and queen and all the people had been turned into stone. One person only had escaped the general transformation, and he was the king's son, who had been brought up as a Muhammadan by his nurse, who was a Musalmani slave. Now this legend appears to be the same as that of

1 Journ. Royal Asiat. Soc, i. 29 and 233.

[p. 301]: Raja Dilu and his brother Chota of the native histories of Sindh, 1 according to which Chota had become a Muhammadan, and when the city of Brahmana was destroyed by an earthquake, on account of the wickedness of the king, Chota alone escaped. As a similar story is told of the ruin of all the chief cities in the Panjab as well as in Sindh, the scene of the story in the ' Arabian Nights ' may be fairly placed in Sindh ; and as Debal was the only large city on the coast, and was besides the chief mart to which the Muhammadan merchants traded, it seems to me almost certain that it must be the Indian city in which Zobeide found all the people turned into stone.

According to M'Murdo, the destruction of Brahmana took place in A.H. 140, or A.D. 757, and as the story of Zobeide is laid in the time of the Khalif Harun-ul-Rashid, who reigned from A.D. 786 to 809, there are no difficulties of chronology to interfere with the identification of the two legends.

The position of Debal may also be fixed on the Baghar river, or main channel of the Indus, by its name of Dibal Sindhi, or Dibal on the Indus. That it was near Lari-bandar we learn incidentally from Captain Hamilton,! who says that the river of Sindhi " is only a small branch of the Indus, which appellation is now lost in this country which it so plentifully waters, and is called Divellee, or Seven mouths." This statement shows that the branch of the Indus leading up to Lari-bandar was called Dibali, or the river of Dibal, so late as A.D. 1699, when visited by Hamilton. That

1 M'Murdo, Journ. Royal Asiat. Soc. i. 28 ; and Postans, Journ. Asiat. Soc. Bengal, rii. 193.

2 ' New Account of East Indies,' i. 130.

[p. 302]: this was the Piti branch of the Indus I infer from its other name of Sindhi, which I take to be the same as the Sinthon Ostium of Ptolemy, or the second mouth of the river, reckoning from the west. As the Piti is one of the mouths of the Baghar river, this position agrees with that which I have already assigned to Dibal, on the concurring testimony of all the previous authorities.

Since Hamilton wrote, Lari-bandar itself has been deserted, and the present port of the western half of the Delta is Dharaja, which is only a few miles to the east of Lari-bandar.

IV. Kachh.

The fourth province of Sindh, in the seventh century, was Kachh, and it was still attached to Sindh in the time of Akbar. It is described by Hwen Thsang as situated at 1600 li, or 267 miles, to the south-west of the capital of Sindh, 1 which at that time was Alor, near Bhakar, on the Indus. This agrees with the details given elsewhere, 1 which make the route as follows : from Alor to Brahmana, 700 li to the south, then to Pitasila 300 li to the south-west, and then to Kachh 700 li to the south ; the whole distance being 1650 li. But the general direction is south, instead of south-west, which agrees with the actual position of Kachh. The province is named O-tien-po-chi-lo, which M. Julien renders as Adhyavakila, or Atyanvakela, but for which no Sanskrit equivalent is offered either by himself or by M Vivien de St. Martin. I think, however, that it may be intended for Audumbatira, or

1 M. Julien's ' Hiouen Thsang,' i. 207, 208. See Map No. IX. 2 Ibid., iii. 170.

[p. 303]: Audumbara, which Professor Lassen gives as the name of the people of Kachh. They are the Odomboerae of Pliny 1 , but there is no trace of this name at the present day.

The province is described as being 5000 li, or 833 miles, in circuit, which is much too great, unless the whole of the Nagar Parkar district to the north of the Ran was included, which is most probable, as this tract has always been considered as a part of Kachh, and still attached to it. Taking its northern boundary as stretching from Umarkot to the neighbourhood of Mount Abu, the whole length of frontier will be upwards of 700 miles. The capital, named Kie-tsi-shi-fa-lo, was 30 li, or 5 miles, in circuit. This name is rendered as Khajiswara by M. Julien, and as Kachchheswara by Professor Lassen. But as the Chinese syllable tse represents the cerebral , I think that tsi must have the same value ; and I would therefore read the whole as Kotiswara, which is the name of a celebrated place of pilgrimage on the western shore of Kachh. That this is the place actually intended is rendered certain by the pilgrim's description of its position, which is said to be on the western frontier of the country close to the river Indus, and to the great ocean. 2 This is a most exact description of the position of the holy Kotesar, which is situated on the western frontier of Kachh, on the bank of the Kori branch of the Indus, and close to the great Indian Ocean. This identification is further supported by the

1 Hist. Nat., vi. 23.

2 M. Julien's ' Hiouen Thsang,' iii. 175 : ....

[p.304]: statement that in the middle of the city there was a famous temple of Siva. The name of the place is derived from Koti+iswara, or the " ten million Iswaras," and refers to the small limgam stones that are found there in great numbers. Iswara is the well known name of Siva, and the lingam is his symbol.

M. Vivien de St. Martin has identified this capital with Karachi ; but the distance from Alor is not more than 1300 li, or 217 miles, while only the initial syllable of the name corresponds with the Chinese transcript. The country is described by Hwen Thsang as low and wet, and the soil impregnated with salt. This is an exact description of the low-lands of Kachh which means a "morass" (Kachchha) , and of the salt desert, or Ran (in Sanskrit Irina), which forms about one-half of the province. But it is quite inaccurate if applied to the dry sandy soil of Karachi. There is also a large swamp extending for many miles, immediately to the south of Kotesar.

Districts to the West of the Indus.

To the west of the Lower Indus all the classical writers agree in placing two barbarous races called Arabii, or Arabitae, and Oritae, or Horitae, both of whom appear to be of Indian origin. The country of the Arabii is said by Arrian to be the "last part of India " towards the west, and Strabo also calls it a " part of India," 1 but both exclude the Oritae. Curtius, however, includes the Horitae in India, 2 while Diodorus states that generally they resemble the

1 Arrian, ' Indica,' 22 ; Strabo, Geogr., xv. 2, 1. 2 Curtius, Vita Alex., ix. 10, 33.

[p.305]: Indians ; and Arrian admits that the Oritae, who " inhabited the inland parts, were clothed in the same manner as the Indians, and used the same weapons, but their language and customs were different." In the seventh century, however, both their language and customs were considered to be like those of the Indians by a much more competent observer, the Chinese pilgrim, Hwen Thsang. According to him, the customs of the inhabitants of Lang-kie-lo, which was 2000 li, or 333 miles, to the west of Kotesar, in Kachh, were like those of the people of Kachh, and their written characters closely resembled those of India, while their language was only slightly different. 1 For these reasons I think that the Oritae, as well as the Arabitae, may fairly be included within the geographical limits of India, although they have always been beyond its political boundary during the historical period. As early as the sixth century B.C. they were tributary to Darius Hystaspes, and they were still subject to Persia nearly twelve centuries later, when visited by Hwen Thsang. But their Indian origin is beyond all doubt, as will be shown when I come to speak of the Oritae.

Arabii, or Arabitae

The Arabii of Arrian are the Arabitae of Curtius, the Arbiti of Ptolemy, the Ambritae of Diodorus, and the Arbies of Strabo. They are said to have derived their name from the river Arabia, or Arbis, or Arabius, which flowed along their confines, and divided their territory from that of the Oritae. 2 From a comparison of the

1 M. Julien's ' Hiouen Thsang,' iii. 177.

2 Arrian, ' Indica, 21 ; Strabo, Geogr., xv. 2. 1 ; Pliny, Hist. Nat., vii. 2.

[p.306]: details of Alexander's marches with the diary of Nearchus, it is certain that this boundary river was the Purali, which flows through the present district of Las into the bay of Sonmiani. According to Curtius, 1 Alexander reached the eastern boundary of the Arabitae in nine days from Patala, and their western boundary in five days more. Now, from Haidarabad to Karachi, the distance is 114 miles, and from Karachi to Sonmiani 50 miles, 2 the former being usually performed by troops in nine marches, and the latter either in four or five. Karachi, therefore, must have been on the eastern frontier of the Arabitae, a deduction which is admitted by the common consent of all inquirers, who have agreed in identifying the Kolaka of Ptolemy and the sandy island of Krokola, where Nearchus tarried with his fleet for one day, with a small island in the Bay of Karachi. Krokola is further described as lying off the mainland of the Arabii. It was 150 stadia, or 17-1/4 miles, from the western mouth of the Indus, which agrees exactly with the relative positions of Karachi and the mouth of the Ghara river, if, as we may fairly assume, the present coast-line has advanced 5 or 6 miles during the twenty-one centuries that have elapsed since the time of Alexander. The identification is confirmed by the fact that "the district in which Karachi is situated is called Karkalla to this day." 3

On leaving Krokola, Nearchus had Mount Eiros (Manora) on his right-hand, and a low flat island on his left, which is a very accurate description of the

1 Vita Alex., ix. 10, 33.

2 Eastwick, 'Handbook of Bombay,' pp. 474 and 477.

3 Ibid., p. 476 ; Burnes, ' Bokhara,' i. 10, writes the name Crocola.

[p. 307]: entrance to Karachi harbour, and after stopping at several small places, reached Morontobara, which was called the " Women's Haven " by the people of the country. 1 From this place he made two courses of 70 stadia and 120 stadia, or altogether not more than 22 miles, to the mouth of the river Arabius, which was the boundary between the country of the Arabii and the Oritae. The name of Morontobara I would identify with Muari, which is now applied to the headland of Ras Muari, or Cape Monz, the last point of the Pabb range of mountains. Bára, or bári, means a roadstead or haven, and moronta is evidently connected with the Persian mard, a man, of which .the feminine is still preserved in Kashmiri, as mahrin, a woman. The haven itself may be looked for between Cape Monz and Sonmitini, but its exact position can not be determined. From the distances given by Arrian in his account of the voyage of Nearchus, I am inclined to fix it at the mouth of the Bahar rivulet, a small stream which falls into the sea about midway between Cape Monz and Sonmiani. If I am right in considering Muari as an abbreviation of Morontobara, the cape must have received its name from the neighbouring haven. At the mouth of the Arabius Nearchus found a large and safe harbour, corresponding with the present Bay of Sonmiani, at the mouth of the Purali, which is described by Pottinger 2 as "a very noble sheet of water, capable of affording anchorage to the largest fleet."

Oritae, or Horitae.

On crossing the river Arabius, Alexander marched

1 Arrian, ' Indica,' p. 22. 2 ' Biluchistan,' p. 9.

[p.308]: for a whole night through a desert, and in the morning entered a well-inhabited country. Then coming to a small river, he pitched his tents, and waited for the main body of the army under Hephsestion. On its arrival, says Arrian, Alexander " penetrated further into the country, and coming to a small village which served the Oritae instead of a capital city, and was named Rambakia, he was pleased with its situation, and imagining that it would rise to be a rich and populous city, if a colony were drawn thither, he committed the care thereof to Hephaestion." 1 On the approach of Alexander, the Oritae made their submission to the conqueror, who appointed Apollo-phanes their governor, and deputed Leonatus with a large force to await the arrival of Nearchus with the fleet, and to look after the peopling of the new city. Shortly after Alexander's departure, the Oritae rose against the Greeks, and Apollophanes, the new governor, was slain, but they were signally defeated by Leonatus, and all their leaders killed. 1 Nearchus places the scene of this defeat at Kokala, on the coast, about halfway between the rivers Arabius and Tomerus. Pliny calls the latter river the Tonberos 2 and states that the country in its neighbourhood was well cultivated.

From these details I would identify the Oritae, or Horitae, or Neoteritae as they are called by Diodorus, with the people on the Aghor river, whom the Greeks would have named Agoritae, or Aoritae, by the suppression of the guttural, of which a trace still remains in the initial aspirate of Horitae. In the bed of this

1 Arrian, Anab., vi. 21, 22 ; and ' Indica,' 23 ; Curtius, ix. 10, 34. 2 Hist. Nat., vi. 25.

[p. 309]: river there are several jets of liquid mud, which, from time immemorial, have been known as Ram-Chandar-ki-kup, or " Ram Chaudar's wells." There are also two natural caves, one dedicated to Kali, and the other to Hingulaj, or Hingula Devi, that is, the “Red Goddess," who is only another form of Kali. But the principal objects of pilgrimage in the Aghor valley are connected with the history of Rama. The pilgrims assemble at the Rambagh, because Rama and Sita are said to have started from this point, and proceed to the Gorakh Tank, where Rama halted; and thence to Tonga-bhera, and on to the point where Rama was obliged to turn back in his attempt to reach Hingulaj with an army. Rambagh I would identify with the Rambakia of Arrian, and Tonga-bhera with the river Tonberos of Pliny, and the Tomerus of Arrian. At Rambakia, therefore, we must look for the site of the city founded by Alexander, which Leonatus was left behind to complete. It seems probable that this is the city which is described by Stephanus of Byzantium as the " sixteenth Alexandria, near the bay of Melane1 Nearchus places the western boundary of the Oritae at a place called Malana, which I take to be the bay of Malan, to the cast of Ras Malan, or Cape Malan of the present day, about twenty miles to the west of the Aghor river. Both Curtius and Diodorus 2 mention the foundation of this city, but they do not give its name. Diodorus, however, adds that it was built on a very favourable

1 In voce Alexandria, <greek>

2 Curtius, Vita Alex., ix. 10: — "In hac quoque regione urbem condidit." Diodorus, Hist. xvii.

[p.310]: site near the sea, but above the reach of the highest tides.

The occurrence of the name of Rambagh at so great a distance to the west of the Indus, and at so early a period as the time of Alexander, is very interesting and important, as it shows not only the wide extension of Hindu influence in ancient times, but also the great antiquity of the story of Rama. It is highly improbable that such a name, with its attendant pilgrimages, could have been imposed on the place after the decay of Hindu influence. 1 During the flourishing period of Buddhism many of the provinces to the west of the Indus adopted the Indian religion, which must have had a powerful influence on the manners and language of the people. But the expedition of Alexander preceded the extension of Buddhism, and I can therefore only attribute the old name of Rambakia to a period anterior to Darius Hystaspes.


These districts are described by Hwen Thsang under the general name of Lang-kie-lo, which M. Julien renders by Langala. M. de St. Martin, however, refers it to the tribe of Langa, but it is extremely doubtful whether this is an ancient name, The other name of Langalas, quoted from the Vishnu Purana, is only a variant reading of Jangalas, which is almost certainly the correct form, as it is immediately followed by Kuru-Jangalas. Hwen Thsang fixes the capital of Lang-kie-lo at 2000 li, or 333 miles, to the west of Kotesar in Kachh. But as this bearing would place it in the middle of the Indian Ocean, the

1 Hingulaj (Khingalatchi) is mentioned by the Tibetan Taranath, see ' Vassilief,' French translation, p. 45, as a Rakshasa in the west of India, beyond Barukacha, or Baroch.

[p.311]: true direction must be north-west. Now this latter bearing and distance correspond with the position of the great ruined city of Lakorian, which Masson1 found between Khozdar and Kilat. In older maps the name is written simply Lakura, which appears to me to be very fairly represented by the Chinese Lang-kie-lo, or Lankara. 2 Masson describes the ruined fortifications as " remarkable for their magnitude, as well as for the solidity and the skill evident in their construction." From the size and importance of these ruins, I conclude that they are the remains of a large city, which has at some former period been the capital of the country. The Chinese pilgrim describes the province as being many thousands of li in breadth as well as in length. It is clear, therefore, that it corresponded, as nearly as possible, with the modern district of Biluchistan, of which the present capital, Kilat, is only 60 miles to the north of Lakura. In the seventh century, the capital was called Su-neu-li-shi-fa-lo, and was 30 li, or 5 miles, in circuit. The Chinese syllables are rendered by M. Julien as Sunuriswara, of which he offers no translation. But as Hwen Thsang describes a magnificent temple of Siva in the middle of the city, I infer that the Chinese transcript may be intended for Sambhuriswara, which is a well-known title of Siva as the "lord of divine beings," or the " god of gods." By assuming that this name belongs properly to the temple, the other name of Lang-kie-lo, or Lakara, may be applied to the capital as well as to the province.

1 ' Kilat,' p. 63 ; and ' Biluchistan,' ii. 46.

2 The same Chinese character, lang, is found in the transcript of Baghalan, where the vowel of the final syllable is long.

Back to Index of the Book