about 50 kos, or say 100 to 125 miles from Soumelpour, which is another detail irreconcileable with the Semah site. The information contained in Tavernier's Voyages soon came to be incorporated by European cartographers in their maps of India. I shall quote here only a few examples out of many. In Mortier's edition of Sanson's atlas, published in 1690-96, we find "Rodas" marked, and about 20 lieues SSE. of it a place called "Soumelpour". On H. Moll's "Map of the East Indies", published a few years later, we find "Rotas" marked, and about 60 miles South of it "Sounielpour" (=Soumelpour) and roughly 20 miles South of that again "Jonpour", both of these places being shown as on a river that flows northwards past Rohtas and joins, a little further on, a shorter river that seems to be intended for the Son. This map by Moll is undated, but must have been published between the years 1701 and 1711. Again on Le Sieur Robert's map of 1741 and on Bellini's map of 1752 we find "Soumelpour" and "Jounpour marked in much the same situations. Now, as none of these names appear on Sanson's map of 1652 and neither Soumelpour nor Jonpour is marked on any of the earlier maps, it is clear that the cartographers had been guided by Tavernier's records. This is manifest in respect of a large number of sites visited by Tavernier during his journeys, which have been shown for the first time on these maps. But as Tavernier makes no mention of any place called Jonpour or Jounpour, it seems to be clear that this name was obtained from some other record then available; and the only traveller whom we can suspect of having furnished this name is L'Escot of Orleans. Now, there is no town or village called Jonpour or Jounpour in this region. Having regard, however, to the not uncommon confusion between capital S and capital J in MSS. the conclusion seems irresistible that this name should have been printed Sonpour or Sounpour. That the initial S had been misread as J is further corroborated by the fact that Rennell, on his map of India published in 1782, marks Sonpur on the Mahanadi as "Sonepour or Jounpour." This solution, it will readily be seen, throws a flood of light on the question of the identification of Tavernier's "SOUMELPOUR", as the two towns marked on the maps cited above are thus obviously no other than Sambalpur and Sonpur-both places of established importance in the middle of the seventeenth century. In fact, Sonpur is most probably the other bourg mentioned by Tavernier as lying 20 kos distant from Soumelpour on the same river. Sonpur is, as a matter of fact, just about 20 kos from Sambalpur, though it lies down stream therefrom, and not above it, as Tavernier has it. The chief difficulty however lies in the name given to the diamond river. In the Voyages it is called Gouel; in the Histoire des Joyaux it is said to have been called Nage, after the name of a large town which was evidently the headquarters of the raja. L'Escot does not appear to have himself published any account of his visit to Sambalpur. The information contained in the Voyages was widely diffused. During the succeeding three-quarters of a century cartographers showed a river Gouel as issuing from mountains lying somewhere to the north of the Mahanadi, and flowing northwards to join the Son near Rohtasgarh; and as the northern Koel does enter the Son in the neighbourhood of Rohtasgarh the Gouel was assumed to represent this Koel. But if Tavernier meant the northern Koel, why should he not have stated that it flowed into the Son, a river with which he was well acquainted, having crossed it between Sasarām and Daudnagar? He does not even say that it flowed northwards. All he says is that it issued from the southern mountains, and flowed on "to lose its name in the Gange." Could the name be intended for the southern Koel? or even for the Ib? Commencing with Lavanha's map of Bengal (drawn about 1550 A.D.), for some two hundred years a river called Ganga or Guenga is shown on the maps as rising in the west of the peninsula and flowing eastwards, to fall into the Bay of Bengal. Lavanha shows it as falling into the Hugli near Tamluk. De Barros describes it as rising in the Western Ghats, and, after flowing through many lands and through the kingdom of Orissa, falling into the Ganges below Satgānv, that is into the Hugli where the Rūpnārāyaṇ enters it. Ramusio (1554) also makes it debouch there. Gastaldi (1561) shows it as having two mouths, one near Hijli and one near Pipli. Mercator, on his world map of 1569, makes it fall into the Hugli near Tamluk. Ortelius (1570) and Thevet (1575) do the same. In the map prepared in 1592, as it seems, to illustrate Linschoten's travels the main channel issues by Pipli, a small channel flowing into the Hugli by Hijli. Hondius, in his early world map (1608), gives it a similar course. Baffin (1619) shows it as debouching by Pipli. Speed (1626) shows two mouths, one near Pipli (=Suvarṇarekhā) and one near Balasore (=Burhabalang). Sanson (1652) also shows two mouths, one near Tamluk (Rūpnārāyaṇ) and one near Pipli (=Suvarṇarekhā). Moll (1701) shows one branch ("R. Sait") as issuing to the north of Balasore and one near Pt. Palmyras (which seems intended for the Brāhmaṇī); but on a later map (!) shows only one mouth near Tamluk. De l'Isle (1739) makes it issue about where the mouth of the Brahmani is; while D'Anville in his map of 1752 makes one branch enter the Bay near Pt. Palmyras and another run round to the north of Balasore and Pipli to the mouth of the Hugli. In fact the courses of the Mahanadi, Brāhmaṇī, Burhabalang, Suvarṇarekha, etc. seem to have been quite unknown down to the middle of the eighteenth century. Whitchurch's map engraved in 1776, but "drawn chiefly from actual surveys, 1769", as he tells us, is the first map I know of that eliminates most of the previous misconceptions, and that shows the course of the "Maha-nudi" from Sambalpur past Sonpur, Baud and Cuttack to the sea, though even on this map the mouth of the river as it enters the Bay lies halfway between Pt. Palmyras and Balasore! The details shown between Sambalpur and Cuttack were clearly obtained from the record of Motte's journey in 1766. It is not impossible that Tavernier's Gange, in which the Gouel lost its name, was meant for this southern Ganga. The possibility of some such confusion is borne out to some extent by the remark made by Mustafa (1758) about the Ib, which he says flows southwards as far as Sambalpur, where it turns eastwards, "and empties itself, I suppose, between Catteck and 4 2 Res. J. " Balassor (I reserve to this River and to the Ganga an extensive Note, De l'Isle seems to have confounded them) ". Unfortunately this promised note, if ever written, is not available to us. Mustafa here appears to refer to De l'Isle's map of India published about 1789. On this map the Mahānadi is marked as Rivière de Ganga ou de Canaca." Immediately to the north of it, and running east and west, are marked "Montagnes du Midi. From these mountains two rivers issue (between 21 and 22 N. Lat.), flowing northwards to join the Son, the western being marked "R. de Gouel où se trouvent les Diamans." On this river, just below 23 N. Lat, are marked "Soumelpour and, close to it, to the SW., "Jounpour". It is particularly noticeable that on this map no rivers are shown as flowing southwards into the Mabana di from these "Montagnes du Midi"; in fact the Ganga (Mahanadi) is shown as flowing past their southern slopes. This part of the map was clearly intended to reproduce the data printed in Tavernier's Voyages. Sufficient has been written perhaps to indicate that Soumel pour was no other than Sambalpur, and that Tavernier himself never went there, but that he probably made a note of what he heard from l'Escot, which was incorporated in the Voyages by his editors in the puzzling form in which we find it. The mistakes are probably due to this. As regards the distance of Soumel pour from Rohtasgarh as given, viz. 30 kos, I can only suggest that kos here is an error, of his editors or printer, for goa. In three places Tavernier explains the gos—a measure of distance which he also uses-as being equivalent to 4 kos. If we read gos, the distance from Rohtasgarh becomes 120 kos, or 120 French lieues, which taking the old lieue at 2:422 English miles, is equivalent to about 290 miles. In a direct line across the map from Rohtasgarh to Sambalpur the distance is 275 miles; and by road it might well be 290 miles, as a very circuitous route would have to be taken. It is just possible that when at Sasarām or in its vicinity, about five years after he had met l'Escot, Tavernier made some inquiry as to the distance of Sambalpur, and was told that it was 30 gos (=gau, Sans. fa: ), or thirty "stages" from Rohtasgaṛh. |