give 6 marches from Charchand to Lob, which is in fair accordance with Polo's 5, and Shaw estimates the whole distance from Ilchi to Lob at 373, or by another calculation at 384, miles, say roundly 380 miles. This higher estimate is to be preferred to Mr. Johnson's for a reason which will appear under next chapter.

Mr. Shaw's informant, Rozi of Khotan, who had lived 12 years at Charchand, described the latter as a small town with a district extending on both sides of a stream which flows to Lob, and which affords Jade. The people are Mussulmans. They grow wheat, Indian corn, pears and apples, &c., but no cotton or rice. It stands in a great plain, but the mountains are not far off. The nature of the products leads Mr. Shaw to think it must stand a good deal higher than Ilchi (4000), perhaps at about 6000 feet. I may observe that the Chinese hydrography of the Kashgar Basin, translated by Julien in the N. An. des Voyages for 1846 (vol. iii.), seems to imply that mountains from the south approach within some 20 miles of the Tarim river, between the longitude of Shayar and Lake Lop. The people of Lob are Mussulman also, but very uncivilised. The Lake is salt. The hydrography calls it about 200 li (say 66 miles) from E. to W. and half that from N. to S., and expresses the old belief that it forms the subterranean source of the HwangHo. Shaw's Itinerary shows "salt pools" at six of the stations between Kiria and Charchand, so Marco's memory in this also was exact.

Nia, a town two marches from Kiria according to Johnson, or four according to Shaw, is probably the ancient city of Ni-jang of the ancient Chinese Itineraries, which lay thirty or forty miles on the China side of Pima, in the middle of a great marsh, and formed the eastern frontier of Khotan bordering on the Desert. (J. R. G. S. XXXVII. pp. 13 and 44; also Sir H. Rawlinson in XLII. p. 503; Erskine's Baber and Humayun, I. 42; Proc. R. G. S. vol. XVI. pp. 244-249; J. A. S. B. IV. 656; H. de la V. de Khotan, u.s.)

CHAPTER XXXIX.

OF THE CITY OF LOP AND THE GREAT DESERT.

Lop is a large town at the edge of the Desert, which is called the Desert of Lop, and is situated between east and north-east. It belongs to the Great Kaan, and the people worship Mahommet. Now, such persons as propose to

cross the Desert take a week's rest in this town to refresh themselves and their cattle; and then they make ready for

the journey, taking with them a month's supply for man and beast. On quitting this city they enter the Desert.

The length of this Desert is so great that 'tis said it would take a year and more to ride from one end of it to the other. And here, where its breadth is least, it takes a month to cross it. "Tis all composed of hills and valleys of sand, and not a thing to eat is to be found on it. But after riding for a day and a night you find fresh water, enough mayhap for some 50 or 100 persons with their beasts, but not for more. And all across the Desert you will find water in like manner, that is to say, in some 28 places altogether you will find good water, but in no great quantity; and in four places also you find brackish water.'

Beasts there are none; for there is nought for them to eat. But there is a marvellous thing related of this Desert, which is that when travellers are on the move by night, and one of them chances to lag behind or to fall asleep or the like, when he tries to gain his company again he will hear spirits talking, and will suppose them to be his comrades. Sometimes the spirits will call him by name; and thus shall a traveller ofttimes be led astray so that he never finds his party. And in this way many have perished. [Sometimes the stray travellers will hear as it were the tramp and hum of a great cavalcade of people away from the real line of road, and taking this to be their own company they will follow the sound; and when day breaks they find that a cheat has been put on them and that they are in an ill plight.] Even in the day time one hears those spirits talking. And sometimes you shall hear the sound of a variety of musical instruments, and still more commonly the sound of drums. [Hence in making this journey 'tis customary for travellers to keep close together. All the animals too have bells at their necks, so that they cannot easily get astray. And at sleeping-time a signal is put up to show the direction of the next march.]

So thus it is that the Desert is crossed.

NOTE 1.-LOP appears to be the Napopo, i.e. Navapa, of Hwen Thsang, called also the country of Leulan, in the Desert (Mém. II. p. 247). Navapa looks like Sanscrit. If so, this carries ancient Indian influence to the verge of the great Gobi. It is difficult to reconcile with our maps the statement of a thirty days' journey across the Desert from Lop to Shachau. Ritter's extracts, indeed, regarding this Desert, show that the constant occurrence of sandhills and deep drifts (our traveller's "hills and valleys of sand") makes the passage extremely difficult for carts and cattle (III. 375). But I suspect that there is some material error in the longitude of Lake Lop as represented in our maps, and that it should be placed something like three degrees more to the westward than we find it (e. g.) in Kiepert's Map of Asia. By that map Khotan is not far short of 600 miles from the western extremity of Lake Lop. By Johnson's Itinerary (including his own journey to Kiria) it is only 338 miles from Ilchi to Lob. Mr. Shaw, as we have seen, gives us a little more, but it is only even then 380. Polo unfortunately omits his usual estimate for the extent of the "Province of Charchan," so he affords us no complete datum. But his distance between Charchan and Lob agrees fairly as we have seen with that both of Johnson and of Shaw, and the elbow on the road from Kiria to Charchand (supra, p. 199) necessitates our still further abridging the longitude between Khotan and Lop. (See Shaw's remarks in Proc. R. G. S. XVI. 243.)

NOTE 2." The waste and desert places of the Earth are, so to speak, the characters which sin has visibly impressed on the outward creation ; its signs and symbols there. . . . Out of a true feeling of this, men have ever conceived of the Wilderness as the haunt of evil spirits. In the old Persian religion Ahriman and his evil Spirits inhabit the steppes and wastes of Turan, to the north of the happy Iran, which stands under the dominion of Ormuzd; exactly as with the Egyptians, the evil Typhon is the Lord of the Libyan sand-wastes, and Osiris of the fertile Egypt." (Archbp. Trench, Studies in the Gospels, p. 7.) Terror, and the seeming absence of a beneficent Providence, are suggestions of the Desert which must have led men to associate it with evil spirits, rather than the figure with which this passage begins; no spontaneous conception surely, however appropriate as a moral image.

"According to the belief of the nations of Central Asia," says I. J. Schmidt," the earth and its interior, as well as the encompassing atmosphere, are filled with Spiritual Beings, which exercise an influence, partly beneficent, partly malignant, on the whole of organic and inorganic nature. . . . . Especially are Deserts and other wild or uninhabited tracts, or regions in which the influences of nature are displayed on a gigantic and terrible scale, regarded as the chief abode or rendezvous of evil Spirits. . . . . And hence the steppes of Turan, and in particular the great sandy Desert of Gobi have been looked on as the dwelling-place of malignant beings, from days of hoar antiquity."

The Chinese historian Matwanlin informs us that there were two roads from China into the Uighur country (towards Karashahr). The longest but easiest road was by Kamul. The other was much shorter, and apparently corresponded, as far as Lop, to that described in this. chapter. "By this you have to cross a plain of sand, extending for more than 100 leagues. You see nothing in any direction but the sky and the sands, without the slightest trace of a road; and travellers find nothing to guide them but the bones of men and beasts and the droppings of camels. During the passage of this wilderness you hear sounds, sometimes of singing, sometimes of wailing; and it has often happened that travellers going aside to see what those sounds might be have strayed from their course and been entirely lost; for they were voices of spirits and goblins. "Tis for these reasons that travellers and merchants often prefer the much longer route by Kamul." (Visdelou, p. 139.)

"In the Desert" (this same desert), says Fahian, "there are a great many evil demons; there are also sirocco winds, which kill all who encounter them. There are no birds or beasts to be seen; but so far as the eye can reach, the route is marked out by the bleached bones of men who have perished in the attempt to cross.'

Hwen Thsang, in his passage of the Desert, both outward and homeward, speaks of visual illusions; such as visions of troops marching and halting with gleaming arms and waving banners, constantly shifting, vanishing and reappearing, "imagery created by demons." A voice behind him calls," Fear not! fear not!" Troubled by these fantasies on one occasion he prays to Kwanin (a Buddhist divinity); still he could not entirely get rid of them; but as soon as he had pronounced a few words from the Prajna (a holy book), they vanished in the twinkling of an eye.

These goblins are not peculiar to the Gobi, though that appears to be their most favoured haunt. The awe of the vast and solitary Desert raises them in all similar localities. Pliny speaks of the phantoms that appear and vanish in the deserts of Africa; Aethicus, the early Christian cosmographer, speaks, though incredulous, of the stories that were told of the voices of singers and revellers in the desert; Mas'udi tells of the Ghuls, which in the deserts appear to travellers by night and in lonely hours; the traveller, taking them for comrades, follows and is led astray. But the wise revile them and the Ghúls vanish. Thus also Apollonius of Tyana and his companions, in a desert near the Indus by moonlight, see an Empusa or Ghúl taking many forms. They revile it, and it goes off uttering shrill cries. Mas'udi also speaks of the mysterious voices heard by lone wayfarers in the Desert, and he gives a rational explanation of them. Ibn Batuta relates a like legend of the Western Sahara : "If the messenger be solitary, the demons sport with him and fascinate him, so that he strays from his course and perishes." The Afghan and Persian wildernesses also have their Ghúl-i-Beában or Goblin of the Waste, a gigantic and fearful spectre which devours travellers; and even

the Gael of the West Highlands have the Direach Ghlinn Eitidh, the Desert Creature of Glen Eiti, which, one-handed, one-eyed, one-legged, seems exactly to answer to the Arabian Nesnás or Empusa. Nicolo Conti in the Chaldaean Desert is aroused at midnight by a great noise, and sees a vast multitude pass by. The merchants tell him that these are demons who are in the habit of traversing the deserts. (Schmidt's San. Setzen, p. 352; V. et V. de H. T. 23, 28, 289; Pliny, VII. 2; Philostratus, Bk. II. ch. iv. ; Prairies d'Or, III. 315, 324; Beal's Fahian; Campbell's Popular Tales of the W. Highlands, IV. 326; I. B. IV. 382; Elphinstone, I. 291; Chodzko's Pop. Poetry of Persia, p. 48; Conti, p. 4.)

The sound of musical instruments, chiefly of drums, is a phenomenon of another class, and is really produced in certain situations among sandhills when the sand is disturbed. A very striking account of a phenomenon of this kind regarded as supernatural is given by Friar Odoric, whose experience I fancy I have traced to the Reg Ruwán or "Flowing Sand" north of Kabul. Besides this celebrated example, which has been described also by the Emperor Baber, I have noted that equally well-known one of the Jibal Nakús, or " Hill of the Bell," in the Sinai Desert; Wadi Hamade, in the vicinity of the same Desert; the Jibal-ul-Thabúl, or Hill of the Drums," between Medina and Mecca ; one on the Island of Eigg, in the Hebrides, discovered by Hugh Miller; one among the Medanos or Sandhills of Arequipa, described to me by Mr. C. Markham; the Bramador or rumbling mountain of Tarapaca; one in hills between the Ulba and the Irtish, in the vicinity of the Altai, called the Almanac Hills, because the sounds are supposed to prognosticate weather-changes; and a remarkable example near Kolberg on the shore of Pomerania. A Chinese narrative of the 10th century mentions the phenomenon as known near Kwachau, on the eastern border of the Lop Desert, under the name of "the Singing Sands ;" and Sir F. Goldsmid has recently made us acquainted with a second Reg Ruwán, on a hill near the Perso-Afghan frontier, a little to the north of Sístán. The place is frequented in pilgrimage. (See Cathay, p. ccxliv, 156, 398; Ritter, II. 204; Aus der Natur, Leipzig, No. 47 [of 1868], p. 752; Remusat, H. de Khotan, p. 74; Proc. R. G. S. XVII. 91.)

CHAPTER XL.

CONCERNING THE GREAT PROVINCE OF TANGUT.

AFTER you have travelled thirty days through the Desert, as I have described, you come to a city called SACHIU, lying between north-east and east; it belongs to the Great