p. 41 Fa-Hian tells us that a journey of four yojanas to the east from Vaisali brought him to "the confluence of the five rivers," and that then, crossing the river (the Ganges) and going south for one yojana, he arrived at Pataliputra in the kingdom of Magadha. There is not, and there can never have been in historical times, any actual meeting of five rivers at or near Patali- putra, i.e. Patna. We can, however, easily understand what Fa-hian meant if we turn to the Buddhist books and observe the special connection of certain rivers mentioned together. In the Vinayapitaka, for instance, we find in the Cullavagga, 9. 1, 3, and 4 (S.B.E., 20. 301f., 304), the Ganga, the Yamuna, the Aciravati, the Sarabhu, and the Mahi, mentioned as "the great rivers"; and in the Milindapanha, 4. 1, 35, we meet again with these rivers and five others, in the following passage (text, 114; S.B.E., 35. 171):-- "There are five hundred rivers which flow down 'from the Himavanta mountain; but of these ten only are reckoned in enumerations of rivers--the Ganga, the Yamuna, the Aciravati, the Sarabhu, the Mahi, the Sindhu, the Sarasvati the Vetravati, the Vitamsa, and the Chandrabhaga --the others not being included in the catalogue because of their intermittent flow of water." This latter passage gives us two groups of rivers, five in each group, which rise in the Himalayas, and we know so many main rivers which have been omitted, and yet cannot have been omitted because their flow is intermittent, that we feel compelled to seek for some more satisfactory reason for an apparently invidious selection. That reason is to be found in the Buddhist system p. 42 of the universe, which has been expounded by Hardy in the opening chapter of his Manual of Buddhism. The following quotations from pages 15 to 17 will suffice to illustrate the point:- "The great forest is in the northern part of Jambudwipa, which, from the southern extremity, gradually increases in height, until it attains an elevation of 500 yojanas, in the mountains of Gandhamadana, Kailasa, Chitrakuta, and others, there being in all 84,000.. These mountains are inhabited by an infinite number of dewas and yakas, and are beautified by 500 rivers, filled with the most delicious water, and by the seven great lakes, among which is the Anotatta-wila. This lake is 800 miles long, and as many broad and deep; and there are four places in it in which the Budhas, Pase-Budhas, rahats, and rishis are accustomed to bathe; and six other places where the dewas from the six inferior heavens bathe. ..... On the four sides of Anotatta are four mouths or doors, whence proceed as many rivers; they are, the lion-mouth, the elephant, the horse, and the bull. The banks of these rivers abound with the animals from which they take their name. The rivers that pass to the north-east and west flow three times round the lake without touching each other, and after passing through countries not inhabited by man, fall into the sea. The river that runs to the south also passes three times round the lake, then rushes from the midst of a rock, and flows in a straight line 60 yojanas.. It then strikes against another rock, and rises into the sky, like a mount of gems 12 miles in size, flows through the sky for the space of 60 jojanas, and strikes against the rock Tiyaggala. This rock it has broken by its immense force; and after this it violently rushes on a further space of 60 yojanas, after which it flows on an inclined plane, strikes and breaks the ponderous Pansu-parwata or Five Mountains, and again passes on 60 yojanas. It then flows 60 yojanas further, through a cave, strikes the four-sided rock Wijja, and is lastly divided into five streams, like five fingers, that are the five great livers (Ganga, Yumuna, Acirawati, Sarabhu, and Mahi), which, after watering Jambudwipa, fall into the sea." Here we have both the five hundred rivers of the Milindapanha and the five rivers of the Vinayapitaka. The latter flow into each other, not all at one place, but consecutively. We must understand "the confluence of the p. 43 five rivers " as denoting the place where the fifth and last of them flows into the already united stream of the other four. And thus we arrive, as will be explained below, at a confluence of five rivers in the neighbourhood of Patna, as stated by Fa-hian. Further, we now see that Fa-hian's expression,'' the confluence of the five rivers," had a specific meaning which was so familiar to the ordinary Ruddlnist of his day as to obviate the necessity of naming the rivers. And it is possible that a further knowledge of the crude but graphic details of the Ruddhist notions of the geography of India, as described in this system of the universe, might help us to elucidate other problems which arise in connection with the records of the travels of the Chinese pilgrims. How far and how long the geographical myth of the origin of the rivers in the Himalayan watershed has misled the natives of India and China, may be realized when we observe that a map of India prepared in Japan in 1710 to illustrate those travels, and reproduced in Julien's Me'moires sur les Contrees Occidentales, shows the rivers issuing from the mouths of animals as mentioned in the passage which I have quoted from Rardy's work. What are "the five rivers" of Fa-hian? The Ganga and Yamuna, the Ganges and the Jamna, are known to everyone; as also is their confluence at Prayaga, Allahabad. The Sarabhu is the modern Sarju or Gogra; it is the Sarabos of Ptolemy, and the Sanskrit Sarayu, on which Valmiki, in Ramayana, i. 5, 6, places the city of Ayodhya. The Aciravati is the Airavati, the modern Rapti; and when we remember that Airavata was the elephant of Indra, we see how it is that this system of five rivers issues from the mouth of an elephant.1 The Rapti (Aciravati) flows into the Sarju, or Gogra (Sarabhu, Sarayu), in lat. 26 15', long. 83 42', between Barhaj in the Gorakhpur district and Dharampur in the Azamgarh district; and the united two rivers flow into the Ganges near Revelganj in the ----------------------------- 1. The Indus (Sindhu) issues from lion's mouth. Hence that river is called Sinh ka bab. the lion's gate or mouth. p. 44 Saran district. 1 Finally, the river which is now known as the Mahi flows actually into the Gandak, the Great Gandak, about half a mile above its junction with the Ganges, but practically into the Ganges, near Sonpur (Sonepore), in the same district. Regarding this latter river, the Mahi, which is not so well known as the others, I must make the following observations. The identity of it was not known to me when I wrote in J.A.S.B., 1900. 74 ff., about the location of Vaisali. The first clue was given to me by a Resident Engineer of the Bengal and North-Western Railway, who reminded me that, on its course from Sewan to Paleza Ghat, the line crosses a river shown in the railway-map as 'Mhye.'(2) And he informed me that, when he was surveying for the alignment, there was trouble with the learned and priestly Brahmans of the locality, who claimed that the name Mahi should be retained, instead of being replaced by that of another river, the Gandaki, which flows into it. But the identity of the Mahi does not rest upon only the information so given to me. The river is mentioned as the "Mahi nadi" in the Statistical Account of Bengal, 11 (1877). 358, as intersecting the Kasmar pargana of the Saran district; and ----------------------------- 1. It's to be noted that at a point about sixteen miles above the confluence of the Gogra and the Rapti, and near a place called Muhoolah, just west of Dohri, in the Azamgarh district, there branches off to the south and south-east a river,.shown in Indian Atlas Sheet No. 103 as 'Surjoo Suddee,' which flows into the Ganges at Ballia. It is not an insigniticant stream; the Gazetteer of the Ballia district tells us, on p. 128, that it is "navigable for large country vessels for five or six months in the vear and for small boats all.the year round." And it may be added that, at the place where this river leaves the Gogra, the latter river has to be strongly embanked and protected by spurs to beep it to its present course. The plans and estimates have passed through me officially. In fact, the people of the Azamgarh and Ballia districts allege that the 'Surjoo Nuddee' runs in the original bed of the Gogra, and it is feared that the latter river may so break its present south hank as to return wholly to its old course. It is not impossible that in ancient times, and in fact in the davs of Fa-hian and Hiuen Tsiang, this Sarjh Nuddee was the real bed of the Gogra; that there was then no stream between Muhoolah-Dohri and Barhaj; and that consequently the Gogra had its confluence with the united Ganges and Jamna at Ballia, and the Rapti had its own separate confluence with the united three streams near Revilganj. I show the Surjoo Nuddee by a dotted line in the annexed sketch-map. 2. This form was evidently borrowed from the same form estab- lished many long years ago in the case of the river Mahi of Western India, which flows into the Gulf of Cambay. p. 45 that though its total course is only some forty to forty-five miles, it's not an insignificnt stream, is shown by the facts there stated, that it's navigable for boats of 600 maunds all the year round and that during the rains boats of all sizes can go up it. Further, the whole course of the Mahi with the name attached throughout is shown in the Bengal Survey Sheets Nos. 83, 84, 113, and 114, of 1902 to 1904.1 The Mahi leaves the Gandak, the Great Gandak, at Sarangpur, 2 in lat. 26 9', long. 84 58', about eleven miles towards the south-east from Dighwa-Dubauli, well known as the find-place of an ancient copper-plate record (I.A., 15. 105). Flowing through a cut in the Saran embankment on the south of the Great Gandak, and passing a large village called Amnaur, it comes eventually to Sitalpur, about nineteen miles on the east of Chhapra. There it receives the waters of the Gandaki. And, the latter river then losing its own name, the two united rivers flow on under the name of the Mahi into the Gandak, or, as said above, practically into the Ganges. Whether the Mahi of the present day is the ancient Mahi-that is, whether its bed marks the original riveris perhaps open to question. My opinion is that the name is an ancient name of the Gandak, the Great Gandak; that the latter river was flowing along its present course in the times of Fa-hian and Hiuen Tsiang; and that in the modern Mahi we have a branch, or an overflow-channel, of the Gandak, by transfer to which the ancient name has been preserved. But, however that may be, it has now been made cl ear that a river known as the Mahi still exists in the exact locality indicated by Fa-hian. I append a small map showing the five rivers of the Buddhists, referred to by Fa-hian. I think it will illustrate how apt is the simile in the passage quoted from Hardy's ----------------------------- 1. In the Indian Atlas Sheet, No. 103 of 1857, with additions to 1895, the name Mahi is not shown, and the course of the river is given under the name of 'Kuthar N (uddee).' 2. There is another Sarangpur exactly 25 miles due west of this, on the Gandaki. p. 46 Manual, which compares them with the five fingers of an out- spread hand--the Ganges below Patna being like the arm. I would add some brief remarks on two other points of importance, which I shall treat more fully on another occasion. The place at which Fa-hian crossed the Ganges to enter the Magadha kingdom--located by him roughly one yojana towards the north from Pataliputra -- is certainly the Paleza-Digha Ghat or crossing, about three and a half miles on the north-west of the western end of Patna. Here the Ganges is held in by steep banks which have probably confined its stream from most ancient times. Fa-hian reached the confluence of the five rivers and the crossing-place of the Ganges by going four yojanas to the east from Vaisali. This locates Vaisali, or some part of it, about four yojanas to the west from Paleza Ghat. With the yojana = 4 6/11 or roughly 4 1/2 miles (see this Journal, 1906, 1012), we have a distance of eighteen miles, which takes us to Cherand, seven miles towards the southeast by east of Chhapra. There are quite sufficient ancient remains at, between, and in the neighbourhood of Chhapra and Cherand to support my opinion, already expressed in .J.A.S.B., 1900. 77 ff., that is the position of Vaisali.