c. 2 in Geographi Græci Minores, Paris, 1882), and is to be sought

in the district watered by the river Diyala, which joins the Tigris near Baghdad. An Armenian Artemita is mentioned by Ptolemy (c. 13, section 21, and c. 8, section 13, edit. Nobbe, Leipzic, 1843).

Schulz tells us that the present village was in his time sometimes called Atramit (he himself writes it Artamit) "par une transposition de lettres qui rappelle un nom fort significatif dans l'ancienne mythologie orientale" (Journal Asiatique, 1840, ser. 3, vol. ix. p. 310). The same traveller was rewarded for his researches in the vicinity by some interesting finds.

In a little valley about 1 1/4 miles west of the village (une demi lieue), and about a hundred paces from the lake (environ une centaine de pas au dessus du lac), among a quantity of blocks of stone, fallen from the hill above, he discovered a cuneiform inscription, engraved upon one of these blocks. Professor Sayce translates this inscription as follows:--"Belonging to Menuas of the mother Taririas, this monument the place of the son of Taririas she has called" (The cuneiform inscriptions of Van, Journal R. Asiatic Society, London, 1882, vol. xiv. p. 529). Dr. Belck, on the other hand, would render it:--"This abode, which belongs to Tarias, daughter of Menuas, is called the palace of Tarias" (Verhandlungen der Berliner Gesellschaft für Anthropologie, 1895, p. 608).

At a little distance from this block Schulz found another inscription, which, owing to exposure to damp, was scarcely distinguishable. He describes it as being engraved on a large stone on the left hand of an ancient aqueduct, built up of several layers of massive stones, several five or six square feet in height. They are irregular in shape and not connected by cement; but are held together by their own weight. The conduit which they enclose is square in form, and of sufficient height and breadth to enable one to stand up inside it. Schulz endeavoured to penetrate within it, but was unable to proceed further than some twenty paces, the passage being obstructed by a large block which had fallen in from the colossal wall (Schulz, op. cit. p. 313).

The hillside above this little valley separates it, he goes on to say, from a kind of upper terrace, over which runs the way from Van to Vostan, among masses of rock, detached from the adjacent heights. Between these rocks flows the Shamiram Su, an artificial channel which has its source some nine leagues south of Van, and which, after passing through the gardens of Artemid, has been conducted to the immediate neighbourhood of the city, where it debouches into the lake. So far as Schulz was able to follow its course, it was nowhere embanked by masonry (ibid. p. 313). It was on this terrace, at a distance of 1 1/4 miles (une demi lieue) from Artemid in a south-westerly direction, immediately by the side of this Shamiram Su, and on the road between Van and Vostan, that he discovered the important inscription which reads according to Professor Sayce:--"To the children of Khaldis the gracious Menuas, the son of Ispuinis, this memorial has selected. Of Menuas the memorial he has named it. To the children of Khaldis, the multitudinous, belonging to Menuas, the king powerful, the king of multitudes, king of the land of Biainas, inhabiting the city of Dhuspas. Menuas says--whoever this tablet carries away, whoever removes the name, whoever with earth destroys, whoever undoes this memorial; may Khaldis, the air god, the sun god, the gods him in public, the name of him, the family of him, the land of him, to fire and water consign" (Sayce, op. cit. pp. 527-28). Messrs. Belck and Lehmann render the word translated by Prof. Sayce as memorial: canal, aqueduct. The rock upon which the inscription was found is known under the name of Kiziltash from its reddish hue.

In the village of Artemid itself Schulz saw the remains of the wall of an ancient edifice on the summit of the cliff. According to Armenian tradition it was formerly a residence of the Armenian kings. Below it he found an ancient conduit (Schulz, ibid. p. 311). I have summarised Schulz's account afresh because Ritter's summary of it (Erdkunde, x. 294) misled me.

The inscription on the Kiziltash has been photographed by M. Müller-Simonis, and reproduced in his book (Du Caucase au Golfe Persique, Paris, 1892, p. 252). Professor Sayce conjectures that these inscriptions served to commemorate the completion of the works connected with the Shamiram Su, and even goes so far as to suggest that the monuments erected by Queen Taririas may have given rise to the traditions about a great queen which in the course of time became transferred to the mythical Semiramis (op. cit. p. 529).

Dr. Belck has within recent years found four more inscriptions in or near Artemid which have been translated by Sayce (Journal R. Asiatic Soc. 1893, pp. 8 seq.). Two are without importance; the remaining two are in the sense of the inscription on the Kiziltash, and are therefore canal inscriptions.

Among the notices of Artemid contained in the works of travellers, a few useful remarks may be gleaned. Shiel (Journey in 1836) tells us that in his time it was a large Armenian village of about 350 houses. Brant (1838) speaks of it as populous, and alludes to the quantity of fruit which was grown there. He approached it from the side of Vostan and the Anguil Su, after crossing which he came upon the Shamiram Su, which he describes as an open canal, supported by a wall in some places. Schulz was impressed by the squalor of the houses; according to him it was peopled half by Armenians and half by Mussulmans; the latter dwelt below the cliff, on the border of the lake (Schulz, op. cit. p. 310). Ussher (before 1865) calls it an Armenian village, and adds--"The flat summit of the rocky hill, on the slope of which the village stood, was surrounded by an ancient wall, built of huge stones laid one upon another without mortar or cement of any kind, and resembling somewhat in appearance Cyclopean remains" (From London to Persepolis, London, 1865, p. 324). Müller-Simonis (op. cit. p. 270) speaks of the "grandes substructions du caractère le plus ancien" which support the Shamiram Su at a certain point three-quarters of an hour on the further side of Artemid, coming from Van.

[117] See Fig. 108, p. 2.

[118] For Mahmudia and a striking photograph of the castle there, see Binder (Au Kurdistan, Paris, 1887, pp. 123 seq.).

[119] The course of the Shamiram Su has been followed and described by Dr. Belck (Zeitschrift für Ethnologie, Berlin, 1892, pp. 137 seq.). It is carried across the Anguil Su or Khoshab by means of a conduit, made of wood, which spans the stream.

[120] Clarified fat or butter, which is generally used for cooking purposes in the East.

[121] I would refer my reader for further information concerning the origin of the patriarch etc. of Akhtamar to Ritter (Erdkunde, vol. x. p. 261), and to the authorities there cited.

[122] An inscription over the door of the narthex is to the effect that it was constructed by Thomas, Katholikos of Akhtamar, in the year of the Armenian era, 1212 (A.D. 1762).

[123] In the geography ascribed to Vardan, a work of the thirteenth century (translated by Saint Martin, Mémoires sur l'Arménie, vol. ii. p. 429), it is said of Akhtamar: "On y trouve l'admirable monastère de la croix bâti par Kagig, roi des Ardzrouniens." According to Chamchean, quoted by Saint Martin (op. cit. vol. i. p. 140), the monastery was founded in A.D. 653 by a prince of the Reshtuni family, named Theodore.

We are informed by Thomas Artsruni (ninth century) that King Gagik brought the stone for building this church all the way from the province of Aghznikh, extending to the Tigris and now comprised within the vilayet of Diarbekr.

[124] Layard, Nineveh and Babylon, London, 1853, pp. 199 and 414. I would ask my reader to compare the illustration of the bronze bull's head with the head in my photograph (Fig. 141) on the right hand or north of the furthest recess of the apse.

[125] But the stone which Layard saw in the portico has probably been removed to the library.

[126] Of these one was circular in form. If it be the same as Sayce's No. XXIX. (op. cit. vol. xiv. p. 537) it is an inscription of Menuas recording a visit to the island.

[127] Ussher (op. cit. p. 332) tells us that the Kurds had carried off many manuscripts which they destroyed from sheer wantonness, using the covers to make soles for their boots.

[128] I will not attempt to explain or reconcile with one another the maps of Kiepert, Cuinet, and Glascott (Journal R.G.S. 1840, vol. x.), and the surveys of Hommaire de Hell (Extrait du Voyage en Turquie, etc., Paris, 1859) and others. Such is the ignorance of one's guides that one cannot do more than question them closely as to the names of villages and put down the information without much confidence in its exactness. What is true of the names of villages is also true of mountains. That portion of the range which lies on the west of Mount Ardos is named Karkar in Kiepert's map; a friend of mine who had travelled in the country knew it under the name of Varkar. I was not made acquainted with either of these names.

[129] Cuinet places the population of Kindirantz at 4064 souls, which is absurd. Nor are there any Jews in the place. His statistics for the caza include 600 gypsies and some Yezidis; but the Kaimakam assured me that 100 was a better figure for the gypsies, while he was not aware of the presence of any Yezidis.

[130] It does not pretend to be more than a very rough sketch plan. It indicates the various mahallas or quarters.

[131] Merchant in Persia, in Italian Travels in Persia, Hakluyt Society, London, 1873, p. 157.

[132] Tavernier, edition of Paris, 1679, vol. i. book iii. p. 303.

[133] Fleurian, Estat présent de l'Arménie, Paris, 1694.

[134] Grammatica e vocabulario della lingua Kurda composti dal P. Maurizio Garzoni, Roma, 1787. See Ritter, Erdkunde, vol. ix. pp. 628, 630.

[135] Abulfeda quotes Ibn Hauqual and Azizi to the effect that Bitlis was a small and prosperous town seven parasangs distant from Akhlat. He adds that in his time it was surrounded by a semi-ruinous wall.

[136] Saint Martin, Mémoires sur l'Arménie, vol. i. p. 103.

[137] Shiel in Journal R.G.S. 1838, vol. viii. p. 73.

[138] Macdonald Kinneir, Journey through Asia Minor, Armenia, and Kurdistan, London, 1818, pp. 393 seq.

[139] Brant in J.R.G.S. vol. x. pp. 379 seq.

[140] Archives of British Consulate at Erzerum.

[141] In detail the figures are: Mussulmans, 27,673; Gregorian Armenians, 15,317; Armenian Catholics, 130; Armenian Protestants, 647; Syrians, 342. Males and females are given separately in the rough census on which these statistics repose; and, owing to the difficulty of access to the women, the latter are always in an apparent minority.

[142] Vital Cuinet (La Turquie d'Asie, Paris, 1892) gravely asserts that there exist 283 scholastic establishments in the vilayet of Bitlis, with 309 teachers and 18,858 pupils of either sex!

[143] Boré, Corr. et Mémoires, Paris, 1840, vol. i. pp. 398, 399.

[144] Rhétoré, Les Missions Catholiques, Paris and Lyons, 1881, pp. 565-567.

[145] The Sources of the Euphrates, in Journal R.G.S. 1895, pp. 173 seq. Mr. Ainsworth conjectures that the water of this well, which he describes as a crater fountain having a basin 220 feet in circumference, comes from Lake Van. I should doubt it. The same careful observer is not quite right in speaking of it as "the source" of the Kara Su. It is no doubt one of the sources, but the Morkh Su, already mentioned, is the first of these westward-flowing streams. For further particulars in regard to the pool of Norshen see Chap. XVIII. p. 317 of the present volume.

[146] I am indebted to the excellent Yusuf, dragoman of the British Consulate at Erzerum and my friend from childhood, for a copy and translation of this inscription: "In the name of God, the merciful and most compassionate, this is the tomb of the great emir, Melik-ul-Umara, Karanlai Agha, who was taken from this place of corruption to the place of mercy and immortality, a Moslem, believer in one God, on the 5th day of Ramazan in the year 689."

[147] My photograph of the belle of Gotni displays such a lack of good features that I must refrain from reproducing it for fear of belying my impression. In its place I offer a picture of one of the best-looking of her less flourishing comrades.

[148] It would probably be safe to say that the Armenian element predominates in the plain proper, and the Kurdish element in the villages bordering upon the plain along the southern border range. Writing in 1838, Consul Brant reported as follows: "In the whole plain of Mush there are not any Mohammedan peasants intermingled with the Armenians: a fact which would clearly point out this country as belonging rather to Armenia than to Kurdistan; indeed the tent-dwelling Kurds are evidently intruders, and the stationary Kurds, it cannot be doubted, belonged originally to the nomad race" (J.R.G.S. 1840, vol. x. p. 347).

[149] I refer the reader with some hesitation to Cuinet's account of this monastery (La Turquie d'Asie, Paris, 1892, vol. ii. p. 584, vilayet de Bitlis). See also Saint Martin, Mémoires sur l'Arménie, vol. ii. pp. 431, 467.

[150] Brant in Journal R.G.S. 1840, vol. x. p. 351. Koch in the forties estimated the population at 1000 Mohammedan and 415 Armenian families, or a total of about 8000 souls (Reise im pontischen Gebirge, etc., Weimar, 1846, p. 405).

[151] Archives of the British Consulate at Erzerum.

[152] For the Catholics of Mush and Mush plain, see Boré (Correspondance et Mémoires, Paris, 1840, vol. i. p. 398), and Smith and Dwight (Missionary Researches in Armenia, London, 1834, p. 429). They have evidently increased in numbers since the time of these writers.

[153] The subject is discussed by Ritter, Erdkunde, vol. x. p. 816.

[154] Saint Martin, Mémoires sur l'Arménie, vol. i. p. 102.

[155] This total is distributed, according to my estimates, as follows:--

Mush-Surb Karapet 25 miles Surb Karapet-Gumgum 28 1/2 ,, Gumgum-Khinis 24 ,, Khinis-Kulli 23 1/2 ,, Kulli-Mejitli 13 ,, Mejitli-Hasan Kala 22 1/2 ,, Hasan Kala-Erzerum 23 ,, ------------- Total 159 1/2 miles.

[156] I should like to refer my reader to Mr. Ainsworth's valuable book (Travels and Researches in Asia Minor, etc., London, 1842, vol. ii. p. 383) for a description of the two interesting old bridges which he found, one spanning the Murad, some distance east of our ford, and the other a former bed of the Kara Su. See also Koch, Reise im pontischen Gebirge, etc., Weimar, 1846, pp. 410, 411.

[157] The head man in a Christian village is called kiaya, and in a Moslem village mukhtar. He is responsible to Government. There is no official chief of agglomerations of villages, like the Russian Pristav.

[158] The accepted average elevation of the plain of Mush appears to be 4200 feet. The readings of my barometers agree fairly well with this figure.

[159] I have already mentioned the presence of gypsies in the caza of Garchigan. I did not meet with any during my first journey.

[160] See the account of Zenobius of Glak as given in the pages of Ritter (Erdkunde, vol. x. pp. 553 seq. and 703), and of Langlois (Collection des historiens de l'Arménie, Paris, 1867, vol. i. pp. 344 seq.). Zenobius is reputed to have been the first bishop of this monastery.

I must add that the work purporting to have been written by Zenobius and called History of Taron, from which Ritter quotes and which is translated in Langlois--and which the monks of Surb Karapet prize so highly--is regarded by modern scholars as a collection of legends made in the eighth or ninth centuries, and valueless as a historical document (see Gelzer, Die Anfänge der armenischen Kirche, in Verhandlungen der könig. Sächs. Gesellschaft der Wiss. zu Leipzig, phil.-hist. Klasse, 1895, p. 123). A much more trustworthy account of the doings of St. Gregory in this neighbourhood is that given in the Agathangelus treatise. I have summarised it in Vol. I. Ch. XVI. pp. 295, 296.

[161] For some account of the doings of all these worthies see the history of John Mamikonean (translated in Langlois, op. cit. vol. i. pp. 361 seq.).

[162] The older names are Glak Vank (from its first abbot), and Innaknean Vank (nine sources).

[163] The moment that I placed my route on my map, I discovered that not the chaoush but my compass had misled me. The direction, as plotted, was quite wrong, as also were the shoots to known landmarks. Happily I was able to fix the position of Dodan with some confidence during my second journey; and the route has been adjusted accordingly. It is evident that the rocks of the plateau behind Surb Karapet must be heavily charged with magnetite.

[164] From Norshen in the east to the passage out of the Murad at Gurgur is a distance of about forty-five miles. Brant, adopting different results, and possibly different measurements, ascribes to the plain of Mush a length of "nearly forty miles" (J.R.G.S. 1840, vol. x. p. 352).

[165] This altitude was ascertained, and the natural features, described with so much hesitation in the present chapter, were elucidated during the second journey (see Ch. XXI.).

[166] Brant, op. cit. pp. 347 seq.

[167] But I must record the fact that the people of Bashkent, when asked the name of their plain, replied, Khinis ova.

[168] I have not reproduced my photograph of Khamur, for a view of which I may refer my reader to Ch. XII. Fig. 177, p. 252.

[169] In Consul Brant's time (1838) Khinis belonged to the pashalik of Mush, and was supposed to contain no more than 130 houses. It is described as "a most wretched town" (op. cit. p. 345).

[170] I have decided, after all, not to reproduce this photograph.

[171] It is interesting to compare Brant's account of Kulli in 1838. His words are:--"It formerly contained a great many Armenian families. I was told that 200 emigrated to Georgia, and only about 15 Mohammedan families now reside among extensive ruins" (op. cit. p. 344). In 1893 the transformation has been completed, and Kulli has become a Kurdish village. The successive steps of the process, which is of general application, may be defined as follows:

1. Emigration or disappearance of Armenians (friends of Turkey make excuses).

2. Lapse into barbarism: enrolment of Hamidiyeh (friends of Turkey exult).

3. Standing nuisance at the doors of Russia (a heavy calm).

4. Russian conquest (Turkey disappears, her friends having preceded her).

[172] See Ritter (Erdkunde, vol. x. pp. 390 seq.), and Brant (op. cit. p. 341). Hamilton (Researches in Asia Minor, etc., London, 1842, vol. i. p. 185) throws doubt upon the popular belief that this and similar castles were built by the Genoese; but I know not upon what foundation he may have based his scepticism.

[173] Which, by the way, is, I believe, made in England out of cloth. Quousque tandem!

[174] I will again cite Brant's account, written in 1838:--"The greater portion of the Armenian peasantry emigrated into Georgia when the Russian army evacuated Turkey, after the peace of Adrianople; in consequence of which emigration, the population of the villages has been much diminished, and there is a great deal of ground uncultivated for want of hands" (op. cit. p. 341).

[175] The season was, it is true, rather exceptional. But it is a noteworthy fact that all these great plains--Mush, Khinis, Pasin--were without snow at this advanced date. Already in March the snow begins to melt.

[176] Brant estimates the distance between Hasan Kala and Erzerum at only eighteen miles (op. cit. p. 341). If he is speaking of the distance by road he makes, I think, a considerable error. My own estimate is twenty-three miles.

[177] The cone of Sheikhjik was visited by Dr. Wagner in the forties and has been described by him at some length (Reise nach Persien, Leipzig, 1852, vol. i. pp. 231 seq.).

[178] Smith and Dwight, Missionary Researches in Armenia, London, 1834, p. 62.

[179] The Jesuit father, Thomas Charles Fleurian (Estat présent de l'Arménie, Paris, 1694, 8vo, p. 81), speaks of Erzerum as "capitale de la haute Arménie sous la domination du Grand Seigneur ... une fort grande ville ... fort peuplée et fort riche; c'est le centre du commerce de tous ces païs-là. Les caravanes qui vont de Perse à Alep, ou à Smirne, ou à Constantinople; ou celles qui viennent de ces mêmes endroits en Perse passent toutes à Erzerom."

[180] Tournefort, Voyage au Levant, Paris, 1717, vol. ii. p. 279, and cp. Schillinger, Persianische und Ost-Indianische Reise, Nürnberg, 1707, 8vo, p. 81. It is a relief to read the warm sentiments of Tournefort towards Mr. Prescot (such was the name of the British agent) in contrast to the verjuice with which our contemporary French travellers think it their duty to steep their pens when speaking of English enterprise or its agents in distant lands. The contrast enables us to measure the difference between the France of Louis XIV. and that of the Presidents.

[181] Brant in Journal R.G.S. 1836, p. 201.

[182] Smith and Dwight in op. cit. p. 64. There were also 645 families of Armenian Catholics and 50 of Greeks. The remainder were Mussulmans.

[183] C. F. Neumann, Geschichte der Uebersiedlung von 40,000 Armeniern welche im Jahre 1828 aus der Persischen Provinz Adebaidschan nach Russland anwanderten (from Russian of S. Glinka), Leipzig, 1834, 8vo.

[184] Brant, loc. cit. It is generally supposed that not less than 60,000 Armenians, headed by their bishop, accompanied the retirement of Paskevich's army.

[185] Smith and Dwight, op. cit. p. 441.

[186] The plain of Erzerum may be said to commence on the west at the village of Titgir.

[187] The excursion to the Dümlü Dagh is a favourite one in summer. The sources of the Kara Su, or Western Euphrates, have been visited and described by Wagner (Reise nach Persien, Leipzig, 1852, vol. i. pp. 237 seq.) and by Strecker (Zeitschrift für Erdkunde, Berlin, 1869, pp. 159 seq.). For a catalogue of the various species of birds found in the marshes of the Kara Su or in the neighbourhood of Erzerum see Curzon, Armenia, London, 1854, chap. x. pp. 143 seq.

[188] For the citadel and old walls of Erzerum the following works may be consulted:--Reyse von Constantinopel, etc., by the Hoch Edelgeborener Herr Heinrich von Poser und Gross-Nedlitz, Jena, 1675, 4o; Tournefort, Voyage au Levant, Paris, 1717, vol. ii. pp. 260 seq.; Morier, Journey through Persia, Armenia, etc., London, 1812, pp. 320 seq.; Macdonald Kinneir, Journey through Asia Minor, Armenia, etc., London, 1818, pp. 366 seq.; Hamilton, Researches in Asia Minor, Pontus and Armenia, London, 1842, vol. i. pp. 178 seq.; Texier, Description de l'Arménie, Paris, 1842, part i. pp. 68 seq.; Koch, Reise im pontischen Gebirge und türkischen Armenien, Weimar, 1846, pp. 274 seq., and Curzon and Wagner in operibus citatis. Koch informs us of a Cufic inscription on the watch-tower in the citadel which was copied by his companion, Dr. Rosen. He adds that it would be published in due course.

[189] Dalyell in Journal R.G.S. 1863, p. 235; Dove, Zeitschrift für Erdkunde, Berlin, 1859, p. 67. The older travellers mention the circumstance that the houses in Erzerum were constructed of wood. Now they are all built of stone.

[190] I would refer my reader to the accounts of Hamilton, Texier, Curzon, Koch and Tozer.

[191] The Merchant in Persia, who travelled in the early part of the sixteenth century, noticed the emblem of an eagle with two heads and two crowns on the buildings of Diarbekr, once the capital of the Ortukids, and mistook it for the imperial arms. See the translation of his work by Charles Grey (Italian Travels in Persia, Hakluyt Society, London, 1873).

[192] Koch, Reise im pontischen Gebirge, etc., Weimar, 1846, p. 284.

[193] Hamilton was informed by his guide that the Chifteh Minareh itself was built by a Sultan of Iran "570 years ago." That was in 1836. The same traveller speaks of a building in Erzerum somewhat resembling Chifteh Minareh but with one minaret only. It seems to be the same as that described by Texier under the name of Mourgo-Serai. I was assured that no such edifice exists at the present day.

[194] Samuel of Ani in Migne, Patrologiæ cursus completus, series Græca, Paris, 1857, vol. xix. p. 706.

[195] See also Vol. I. Ch. XVI. p. 261.

[196] Procopius, de bell. Pers. lib. i. c. 10. The student must be careful to distinguish this Theodosiopolis from the fortress of the same name on the Khabur. The letter of the emperor to the patriarch Isaac is given by Moses of Khorene, lib. iii. c. 57.

[197] Moses of Khorene, lib. iii. c. 59. Thousands of eggs are still collected in these marshes during spring by the inhabitants of the plain of Erzerum. The hot springs mentioned are evidently those of Ilija, a good hour's drive to the west of Erzerum.

[198] Nöldeke, article "Persia" in Ency. Brit. 9th edit. vol. xviii. p. 611; Procopius, de Edificiis, iii. c. 5.

[199] Procopius in loc. cit. In the time of Justinian the frontier of Roman Armenia skirted the Persian frontier from the city of Amida (Diarbekr) as far as Theodosiopolis (ibid. iii. c. 1).

[200] Indgidgean, ap. Neumann, quoted by Ritter, Erdkunde, vol. x. p. 759.

[201] Issaverdens, Armenia and the Armenians, Venice, 1878, p. 109.

[202] Indgidgean in op. cit.

[203] Cedrenus, edit. Bekker, p. 463; see Gibbon, Decline and Fall, ch. liv.

[204] Constantine Porphyrogenitus, de Adm. Imp. c. 45.

[205] Von Hammer, Geschichte des Osm. Reiches, vol. i. p. 25.

[206] Kyriakos, ap. Ritter, Erdkunde, vol. x. p. 760.

[207] Travels of Evliya, translated by Von Hammer, London, 1850, vol. ii. p. 108.

[208] Abulfeda, Annales, edit. Reiske, iv. p. 367. For a plan and account of the ruins of the southern Arzen see Taylor in J.R.G.S. vol. xxxv. pp. 26 seq. Evliya speaks of four towns bearing the name of Erzen, viz. Erzen in Mesopotamia, Erzen Akhlat, Erzen Rum, commonly called Erzerum, and Erzenjan (Von Hammer's translation, ii. 202). The word Erzen or Arzen is discussed by Boré, Corr. et Mémoires, Paris, 1840, vol. i. pp. 184 seq. Strecker (Zeitschrift für Erdkunde, Berlin, 1869, pp. 152, 153) seeks to identify our Artze or Artsn with the site of the modern village of Karars near the right bank of the Kara Su or Euphrates, north-west of Erzerum.

[209] Cedrenus, pp. 772, 773. He speaks of Artze as a kômopolis in the neighbourhood of Theodosiopolis which is described as a strong fortress. A vivid contemporary and native account of the sack of Artze is furnished by the Armenian historian, Aristakes of Lastivert. See Prudhomme's translation in the Revue de l'Orient, Paris, 1863, vol. xvii. pp. 275 seq.

[210] Saint Martin, Mémoires sur l'Arménie, vol. i. p. 68; Tournefort, Voyage du Levant, Paris, 1717, vol. ii. p. 276.

[211] Erzerum is also known to Armenian writers under the name of Karnoy Kaghak (Kalak) or town of Karin, from which the name of Kalikala, used by Arabic authors, is probably derived.

D'Anville is certainly in error when he seeks to identify Theodosiopolis with Hasan Kala in Pasin.

[212] Since writing this description, General Sir Charles Wilson's most admirable Handbook for Asia Minor (London, Murray, 1895) has come into my hands. He gives the distance between Erzerum and Trebizond, measured in miles along the chaussée, at 199 1/4 miles. Another account makes the total 196 1/2 miles. I enquired in official circles at Erzerum whether there were extant any exact record of the distance; a search was made in the archives with a negative result. A certain proportion of the milestones are still erect; but many have disappeared, the course of the road has been changed in places, and the milestones have been replaced, probably in an arbitrary manner. My own record, which is based on careful estimates of pace and time, is as follows:--Erzerum--Ashkala, 33 miles; Ashkala--Pirnakapan, 10 miles; Pirnakapan--Southern Kop Khan, 2 miles; Southern Kop Khan--Kop Pass, 5 1/2 miles; Kop Pass--Northern Kop Khan, 5 1/3 miles; Northern Kop Khan--Maden Khan, 6 1/2 miles; Maden Khan--Baiburt, 10 3/4 miles; Baiburt (bridge)--Varzahan, 6 miles; Varzahan--Osluk Khan, 6 miles; Osluk Khan--Khadrak, 8 miles; Khadrak--Vavuk Pass, 4 1/2 miles; Vavuk Pass--Murad Khan, 10 1/3 miles; Murad Khan--Lower Gümüshkhaneh, 16 1/4 miles; Lower Gümüshkhaneh--Ardasa, 16 1/2 miles; Ardasa--Southern Zigana hamlet, 9 1/2 miles; Southern Zigana village--Zigana Pass, 4 1/2 miles; Zigana Pass--Upper Hamsi Keui, 10 1/8 miles; Upper Hamsi Keui--Jevizlik, 15 1/4 miles; Jevizlik--Trebizond, 20 miles. Total, 199 3/4 miles. A carriage (victoria) can be obtained in Trebizond. Such a vehicle, drawn by two horses, together with a cart for the luggage with a team of three, costs for the whole journey £13 : 10s. But, if I may offer a recommendation to the traveller, it is to render himself independent of the chaussée by purchasing horses and riding. Large deductions from the mileage may be made in this way, and the jolting avoided which is inseparable from a metalled road kept in bad repair. Indeed wheeled traffic is as yet quite an anomaly both in Turkey and in Persia. On the other hand, it is extremely difficult to buy good horses in Trebizond, although they may be readily purchased in Erzerum.

[213] According to Strecker (Zeit. Erdk. Berlin, 1869, vol. iv. p. 147) the Serchemeh Chai has a shorter course and brings less water than the Kara Su. I should consider that of these two uppermost constituents of the Frat, the former has the greater average volume.

[214] The basin of Ashkala has been treated in its geological aspects by Abich in his usually masterly manner (Geologische Forschungen in den kaukasischen Ländern, Vienna, 1882, pt. ii. sect. 1, pp. 100 seq.).

[215] Macdonald Kinneir (Journey through Asia Minor, etc., London, 1818, p. 358) seems to have mistaken this Terjan range for that on the south of the Murad. He is respectfully followed by the laborious Ritter (Erdkunde, vol. x. p. 743). But that erudite geographer, to whom we owe so much, should have been more careful to qualify the statement (p. 741) that the range which is crossed by the Kop Pass constitutes the "Nordbegrenzung des armenischen Plateaulandes." A few months' personal travel would have stood him in good stead after all his minute analysis of the works of travellers.

[216] But one of them has already been drawn by Layard (Nineveh and Babylon, p. 7); and I here reproduce a photograph taken during our second journey, which shows some interesting examples of old Armenian tombstones with rams' heads in the cemetery of Varzahan (Fig. 173).

[217] For the stages see Ch. XI. p. 240. The route is shown on my map.

[218] Not that all the people one meets of distinctively Greek type are Christians. Especially in the valleys most remote from the coast, as in that of the Kharshut, the inhabitants of Greek race have largely been converted to Mohammedanism, or have become Mohammedan for prudential or worldly motives. So complete has been the transformation in some places, that, when I asked my host at Besh Kilisa--a man whose physiognomy showed him to be a typical Greek--to what nationality he belonged, he replied "Osmanli." A section of the inhabitants of Hamsi Keui--a village south of the Zigana--represent a transitional stage. Their children are baptized, but a mollah recites prayers over them. They bear a Mohammedan and a Christian name, as, for instance, that of Ahmed Apostolos. When they die the papa and the mollah dispute the corpse. They have neither church nor mosque. When they meet a Greek they bid him kallispera, and, when a Mohammedan, merhaba.

[219] The following are the stages:--Jevizlik--Sumelas, 10 1/2 miles; Sumelas across the Kazikly Dagh to Tashköpri, 11 miles; Tashköpri viâ Tshörak Khan and across the Kitowa Dagh to Mezere Khan, 18 1/4 miles; Mezere Khan to Baiburt, 17 1/2 miles. From Baiburt the summer road to Erzerum viâ the Khosabpunar Pass may be taken, the stages being:--Baiburt--Maden Khan, 10 3/4 miles; Maden Khan to Khosabpunar village on the south side of the pass (8600 feet), 28 miles; Khosabpunar village viâ Maimansur to Erzerum, 29 miles. Total distance from Trebizond by this route, 145 miles, as against the 199 miles of the chaussée. See Ch. X. p. 225.

[220] The following are my estimates of the mileage distances along our route to Khinis: Erzerum--Palandöken Pass, 7 1/4 miles; Palandöken Pass--Madrak, 8 miles; Madrak--Khedonun, 11 1/2 miles; Khedonun--Kherbesor, 8 3/4 miles; Kherbesor--Ali Mur, 7 miles; Ali Mur--Khinis, 18 miles. Total, 60 1/2 miles. Such estimates throughout this work are based on pace and on time occupied; and the results have been checked by the positions fixed by cross bearings.

[221] The illustration was taken from a hill near Khedonun, almost in the centre of the basin-like area. But the appearance both of basin and of mountain are substantially the same from the Palandöken Pass. I may refer my reader to the similar landscape taken in winter (Ch. VIII. Fig. 161).

[222] See Ch. VIII. p. 188.

[223] The following were our stages to Tutakh:--Khinis--Dedeveren, 17 miles; Dedeveren--Gunduz, 8 miles; Gunduz--Gopal, 9 miles; Gopal--Rashan, 8 1/4 miles; Rashan--Alkhes, 23 1/2 miles, several of which might have been saved; Alkhes--Tutakh, 18 1/2 miles. Total, 84 1/2 miles.

[224] See Ch. VIII. p. 178.

[225] The subject is discussed, Vol. I. Ch. XXI. pp. 422 seq.

[226] The stages are as follows:--Tutakh--Gargalik, 12 1/2 miles; Gargalik--Melazkert, 24 3/4 miles. Total, 37 1/4 miles.

[227] The calculation is based on the difference between the level of the lake (5637 feet) and that of the Murad at the old bridge of Melazkert (5174 feet). Both levels were taken with the boiling-point apparatus.

[228] The fall between Tutakh and Melazkert can only amount to about 100 feet.

[229] This Yusuf Pasha is not the same as the Yusuf Bey who received me at Köshk (see Ch. II. p. 16).

[230] The original name is Manazkert, which the Turks have corrupted into Melazkert. In the older name there perhaps lurks that of Menuas, the Vannic king, who reigned in the ninth century before Christ (see Ch. IV. p. 71).

[231] Saint Martin, Mémoires sur l'Arménie, vol. i. p. 251.

[232] I have transcribed my impressions, as written on the spot. But it is possible that the present aspect of the walls as well as the bridge may be due to the Mohammedan rulers of Melazkert. The great tower in the citadel may well be later than the eleventh century. Still there can be little doubt that the work was carried out by Armenians, and in harmony with the original plan.

Unfortunately almost all the inscriptions have disappeared. We observed a slab of calcareous stone inserted in the north wall, and engraved with an Arabic inscription, but it was much obliterated. A slab of the same material, and in the same condition, containing an inscription, probably in the Syriac character, is built into the Kaimakam's house. We were told that a number of inscriptions had been abstracted by the son-in-law of Raouf Pasha, Vali of Erzerum.

Outside the citadel, lying upon the ground, we examined a well-preserved cuneiform inscription, engraved upon two sides of a block of granitic rock, unlike any stone found here. I was under the impression that it had already been discovered and translated; so we did not take a copy. I now find that we should have done well to copy it. Scheil (Recueil de travaux relatifs à la philologie et à l'archéologie égypt. et assyr., Paris, 1896, vol. xviii. pp. 75-77) describes and translates an inscription which, he says, was recently discovered at Melazkert by the district engineer, but he does not mention the exact locality. It is an inscription of Menuas, recording a restoration.

Within the citadel, near our encampment, one of those large stones which have been elsewhere described, incised with the elaborate traceries of an Armenian cross, was seen among the debris. It was in excellent preservation, having only recently been dug out in situ.

[233] Our stages were:--Melazkert--Demian, 10 miles; Demian--Akhlat (Erkizan), 20 1/2 miles.

[234] Coracias garrulus, belonging to a family closely allied to the kingfishers and bee-eaters. But what hideous names have been given to this beautiful bird!

[235] The credit of whatever information we already possess is due, among modern travellers, almost exclusively to Englishmen. I may cite Brant (Journal R.G.S. 1840, vol. x. pp. 406 seq.), Layard (Discoveries in the Ruins of Nineveh and Babylon, London, 1853, pp. 24 seq.), and Tozer (Turkish Armenia, London, 1881, pp. 315 seq.). The last of these writers does not appear to have read Layard's account, which would have saved him some lengthy speculations. Ritter (Erdkunde, vol. x. p. 326) may also be consulted.

[236] The white stone which may be seen inserted in the masonry of some of the tombs at Akhlat is not a true marble, but a compact limestone, easy to chisel. It must have been brought from a distance, perhaps from the opposite shore of the lake, as we met with no such stone in situ during our wanderings.

[237] I am indebted to my friend, Mr. E. Denison Ross, for careful translations of these and the following inscriptions.

[238] Brant (op. cit. p. 407) attests its existence at the time of his visit.

[239] Woodcuts of this tomb are given by Layard (op. cit. p. 24) and by Müller-Simonis (op. cit. p. 313).

[240] Madavantz belongs to the caza but not to the casaba, or home district, of Akhlat.

[241] Geography, attributed to Vardan ap. Saint Martin, Mémoires sur l'Arménie, vol. ii. p. 429. One of these monasteries contained the leather girdle of St. Gregory, and another was consecrated by the saint himself.

[242] Constantine Porphyrogenitus, de Adm. Imp. c. 44, in vol. iii. p. 196 of the Bonn edition.

[243] Lane-Poole, Mohammedan Dynasties, London, 1894, p. 118.

[244] Deguignes, Hist. des Huns, Paris, 1756, vol. i. p. 253; Lane-Poole, op. cit. p. 170.

[245] Saint Martin, quoting Chamchean, Hist. vol. iii. p. 221.

[246] Layard (op. cit. p. 26) mentions a local tradition that all these tombs were built by Sultans of the Ak-Kuyunli (White Sheep) and Kara-Kuyunli (Black Sheep) Turkomans. The inscriptions show that this cannot be the case. The Venetian traveller Barbaro, who visited the country during the first half of the fifteenth century, found it in the possession of the horde of the Black Sheep. They were driven out by the rival horde of the White Sheep under Uzun Hasan (1466-1478).

Layard speaks of Bayindar as a known sultan of the White Sheep horde, I know not upon what authority.

[247] Von Hammer, Geschichte des osmanischen Reiches, vol. iii. p. 143. Akhlat appears to have contained the tombs of some of the ancestors of the Ottoman ruling House (ibid. note to p. 144 on p. 676).

[248] The Merchant in Persia (Travels of Italians in Persia, Hakluyt Society, London, 1873, p. 160), who visited Armenia in the first quarter of the sixteenth century, describes it as follows:--"This Calata (sic) was anciently a large city, as can be seen by the buildings, but is now reduced to a small fortress."

[249] The accuracy of the results obtained with our Steward telemeter was well tested on Nimrud by these cross-readings from salient points along the edge of the crater. The principal credit, however, for the excellent measurements, taken under great difficulties, owing to the uneven surface of the ground, is due to Captain Elliot, who was ably seconded by Mr. Oswald.

[250] Among craters similar in character to Nimrud the best comparison would seem to be afforded by the Crater Lake of Oregon (Cascade range). The average diameter of this crater is a little more than 5 1/2 miles, and while the inner slopes around the lake are precipitous, those facing outwards to the platform upon which the crater is reared are gentle. The highest point is 8200 feet above the sea, and the lake has a depth of 2000 feet (see J. S. Diller, American Journal of Science, 1897, p. 165). The well-known craters in the Sandwich Islands are much smaller than Nimrud, the largest, Kilauea, having a maximum diameter of 2 1/2 miles.

[251] Or it may merely represent the terminal walls of lava streams.

[252] A single paragraph in an article by Major Clayton, R.A., entitled "The Mountains of Kurdistan" (Alpine Journal, Aug. 1887), is the only account known to me of the interior of the Nimrud crater. Brant confines himself to the following grotesque description (Journal R.G.S. 1840, vol. x. p. 378):--"The Nimrud range (sic) runs nearly north and south, but at its southern extremity is terminated by a cross range (sic), called the Kerkú Tágh, running east and west."

[253] Merchant in Persia, Travels of Italians in Persia, Hakluyt Society, London, 1873, p. 159.

[254] The name Rava is sometimes applied to this plateau.

[255] The Kaimakam of Akhlat, who knows the district well, assured us that there was a permanent outlet. Layard, on the other hand, speaks of an intermittent one (Discoveries in the Ruins of Nineveh and Babylon, London, 1853, p. 21). I regret that I am unable to express certainty on the point. In the case of Lake Bulama, however, I am able to vouch for the fact that its waters find their way to the Murad.

[256] The following are the approximate distances along the route described in this chapter:--Tadvan to Norshen, 12 3/4 miles; Norshen to Nazik, 23 1/2 miles; Akhlat to Adeljivas, 15 miles.

[257] Oswald took several careful observations of the dip of these limestones. The norm was 50° south by east.

[258] The best account of Adeljivas is that of Tozer (Turkish Armenia, London, 1881, p. 335). The width of the enclosure is given by him as 250 yards, on the side of the shore. The parallel lines of walls descend into the water. Within the enclosure "one ancient mosque with a minaret remains, and also part of another considerable building. The mosque, which is now used as a storehouse for corn, appears to be of the same date as those in the castle of Akhlat; it is massively built of stone, with but little ornament, and its arches are pointed and slightly ogived." Müller-Simonis has a nice woodcut (Du Caucase au Golfe Persique, Paris, 1892, p. 301).

[259] This is the height of the village of Uran Gazi.

[260] It would require a series of very careful observations to determine whether this eminence--which I shall call the eastern summit--or the western summit of Sipan be the higher. By boiling-point we obtained the following results:--Eastern summit (4th August), 13,590 feet; western summit (5th August), 13,714 feet. But these readings were taken on different days. On the other hand, the aneroid registered:--Eastern summit (4th August), 13,650 feet; eastern summit (5th August), 13,790 feet; western summit (5th August), 13,754 feet. At present the question must be left open--and indeed it is not of much importance.

[261] Such ziarets exist upon almost all the prominent mountains, great or small, in this part of Armenia. The custom no doubt comes down from an epoch of Nature-worship.

[262] It will be recognised from the above description that the summit of Sipan is much more basin-like than that of Ararat. Sipan probably possessed a crater in the proper sense. That of Ararat is so much worn down that it can scarcely be said to exist.

Sipan appears to have been built up by successive lava streams, which became more and more viscous, until that finally emitted had no power to flow at all, and merely welled up, forming the circular mass on the east. The lava composing that mass is spongy and glassy, a glassy mica-andesite. The narrow ridge, upon which we camped, and which may represent the northern rim of the old crater, consists of a slabby rhyolite with impure obsidian; it is covered up with cindery slag. The western summit and surrounding rock is made up of a lava somewhat similar to that on Nimrud--a dull impure obsidian with ill-developed spherulites; the flow structure is well marked. Tuffs were nowhere to be seen. But a bastion on the northern side of the mountain was cloaked with grey pumice sand.

The ascent of the mountain is described by Brant (J.R.G.S. 1840, vol. x. pp. 409 seq.) and by Tozer (Turkish Armenia, pp. 327 seq.). But they both largely underestimate the height. They appear to have been misled by the fact that the highest points are free from snow in midsummer; but the summit region in general is a mass of snow even at that season. On Ararat such piles of rock, on which the snow has been unable to obtain a footing, are found quite near the summit, which is nearly 17,000 feet high.

[263] The list is divided into cazas and villages:--Adeljivas caza--1, Uran Gazi; 2, Kogus. Van caza--3, Shikhare; 4, Shikhuna; 5, Azikare; 6, Pakis. Akhlat caza--7, Kholik; 8, Agjavireh; 9, Yogurtyemes; 10, Develik; 11, Khanik. Melazkert caza--12, Serdut; 13, Yarelmish; 14, Kara Ali; 15, Simu. Bulanik caza--16, Gopo. Khinis caza--17, Lekbudagh. Varto caza--18, Charbahur; 19, Charbahur Tepe; 20, Akhpoghan; 21, Zirnek; 22, Budag; 23, Shekan; 24, Aineh. In addition to these--I will not vouch for the spelling--there were, he said, to be found Circassians on the side of Erzerum.

[264] The intermediate distances along the route described in this chapter were as follows:--Uran Gazi viâ Leter and Lake Bulama to Gop, 32 miles; Gop to Charbahur, deviating to the confluence of the Bingöl Su (Khinis) with the Murad, 52 3/4 miles; Charbahur to Gumgum, 6 3/4 miles; Gumgum to Gundemir, 9 1/4 miles. Total, 100 3/4 miles.

[265] Mytilus (Congeria) polymorphus.

[266] Layard (Discoveries in the Ruins of Nineveh and Babylon, London, 1853, p. 20) speaks of it under the name of the Lake of Shailu.

[267] The distances are as follows:--Gop to Karaogli (including one considerable deviation), 12 3/4 miles; Karaogli to Shakhberat, 10 miles; Shakhberat to Charbahur, 30 miles. Total, 52 3/4 miles.

[268] A horse's knee would represent a depth of 1 foot 5 inches, and a horse's girth 2 feet 9 inches.

[269] The lava may be described as a fine-grained augite-andesite, grey in colour with distinct augite crystals. It is slightly scoriaceous superficially.

[270] Existing literature on the subject is not satisfactory. I may cite the following:--Koch, Reise im pontischen Gebirge, etc., Weimar, 1846, pp. 365 seq., and p. 333; Der Kaukasus, Landschafts- und Lebens-Bilder, by the same author, published posthumously, Berlin, 1882. See the chapter entitled "Der Berg der tausend Seen." P. de Tchihatchef (1858), Asie Mineure, part iv., Geology, Paris, 1867, vol. i. pp. 279-285; Kotschy, Reise von Trapezunt, etc., in Petermann's Mittheilungen, 1860; Strecker, Beiträge zur Geographie von Hoch-Armenien, in Zeitschrift der Gesellschaft für Erdkunde zu Berlin, 1869, pp. 512 seq.; Radde in Petermann, 1877, pp. 411 seq. Of these, Radde's article is the most reliable, and is, indeed, a valuable contribution, so far as it goes. Abich has endeavoured to make the best of these accounts. See Geologische Forschungen in den kaukasischen Ländern, Vienna, 1882, part ii. sec. 1, p. 77, and pp. 87 seq.

[271] Strecker, op. cit. p. 516. This writer calls the western summit Toprak Kala, or the earth castle.

[272] A fairly compact augite-andesite.

[273] W. Gifford Palgrave, in Nature, vol. v. 1871-72, p. 444; and vol. vi. 1872, pp. 536 seq.

[274] Strecker (op. cit. p. 515) states that he found a stone three feet long and two feet broad, inscribed with cuneiform characters, lying on the ground in a depression east of Kara Kala. It was surrounded by the gravestones of a little cemetery. For an account of the inscription on our stone see Ch. IV. p. 73.

[275] See Geologische Forschungen in den kaukasischen Ländern, Vienna, 1882, vol. ii. pp. 7, 85, and 89.

[276] Both Oswald and myself had read Abich's account of this so-called crater. He appears to regard it as a volcanic crater in the strict sense. I am inclined to think that his drawing is very much exaggerated (Geologische Forschungen in den kaukasischen Ländern, Vienna, 1882, II. Theil, pp. 73 et seq.).

[277] For the stages see Ch. XI. p. 240.

[278] An account of this route which I have before me gives the distance between Rizeh and Erzerum as only 119 miles. It leaves Ispir (in the Chorokh valley) a little to the east.

[279] I must not omit to record the assistance which I have received from the map of H. Kiepert, Provinces Asiatiques de l'Empire Ottoman. The sheets which cover the Armenian country embody the results of my predecessors, which have been compiled with great judgment. I have also had access to two Russian maps embracing portions of the country, (1) scale 10 versts = one inch, 1889, (2) scale 20 versts = one inch, 1899. But the map of Kiepert with all its merit is necessarily sketchy; and the last Russian map is flagrantly incorrect.

[280] See the map of Loftus in Quarterly Journal of the Geological Society, vol. xi. London, 1855, p. 247.

[281] See the map of Abich in Geologische Forschungen in den kaukasischen Ländern, Vienna, 1882, Atlas, Karte I.; and part ii. p. 141.

[282] A fine view of the range at this point is displayed by Abich, op. cit., Atlas table iii.

[283] Abich, op. cit. part ii. p. 155.

[284] Abich, op. cit. part ii. p. 160.

[285] E. Clayton, The Mountains of Kurdistan, in the Alpine Journal, 1887.

[286] For some account of the geology of Azerbaijan see C. Grewingk, Die geognostischen und orographischen Verhältnisse des nördlichen Persiens, St. Petersburg, 1853.

[287] The best account of this country is that of J. G. Taylor, J.R.G.S. vol. xxxviii. 1868. I may also refer my reader to two articles by Dr. Butyka (Mitt. der K. K. geographischen Gesellschaft in Wien, vol. xxxv. 1892, pp. 99-126 and 194-210), who has collected the scanty notices of his predecessors and added his own experiences. I have made use of some unpublished material in the preparation of this part of my map; but it is far from satisfactory.

[288] See the inscription and translation by Professor Sayce in the Journal of the Royal Asiatic Society, vol. xiv. 1882, No. XXXIII. p. 558.

[289] Tozer, Turkish Armenia and Eastern Asia Minor, London, 1881, pp. 256 seq.

[290] H. Binder describes the route from Mosul through Amadia, Julamerik, Kochannes and Mervanen to Van (Au Kurdistan, en Mésopotamie et en Perse, Paris, 1887). See also W. F. Ainsworth, Travels in Asia Minor, etc., London, 1842, vol. ii. pp. 179 seq., with geological section from Mosul to Lake Urmi.

[291] For the geology of the Taurus between Diarbekr and Kharput an article by W. Warington Smyth in the Quarterly Journal of the Geological Society of London, 1845, vol. i. pp. 330 seq., may be consulted. I do not understand his statement that the breadth of the main ridge of Taurus between Arghana and Kharput is nearly 50 miles. Loftus (op. cit. p. 344) has drawn a geological section from Bitlis through Sert to Jezireh-ibn-Omar. Like the Zagros, the range may be said to consist in the main of nummulitic limestone.

[292] Abich, op. cit. part ii. p. 119 note.

[293] I cannot speak with certainty as to the geological nature of the Pir Reshid Dagh.

[294] For Tendurek, which appears to be in a solfataric condition, see Abich's article in the Bulletin of the French Geological Society, 2nd series, xxi. pp. 213 seq., and Letter from T. K. Lynch in P.R.G.S. xiii. pp. 243, 244.

[295] The Miocene deposits are found in the valleys, e.g. in those of the Frat (Western Euphrates) and Araxes.

An interesting fact has been brought to my notice by Mr. F. Oswald, my friend and companion during my last journey. There may be seen in the Tiflis Museum the remains of a mammoth which was discovered in the lacustrine deposits of the Alexandropol district. Similar remains had already been found in deposits of similar character and age in the neighbourhood of Khinis by Colonel J. Shiel. These are in the British Museum, where they have been christened Elephas armeniacus.

[296] See Blue-book, Turkey, No. 6, 1881, p. 127.

[297] The clause in the Berlin Treaty relating to the Armenians is as follows:--Article 61. "La sublime Porte s'engage à réaliser sans plus de retard les améliorations et les réformes qu'exigent les besoins locaux dans les provinces habitées par les Arméniens et à garantir leur sécurité contre les Circassiens et les Kurdes. Elle donnera connaissance périodiquement des mesures prises à cet effet aux puissances qui en surveilleront l'application."

The Armenian delegates to the Berlin Congress presented a memoir to the European Plenipotentiaries in which they set forth their cause. It is published by De la Jonquière in his Histoire de l'Empire Ottoman, Paris, 1881, pp. 39-44. They also drew out a project for an Organic Regulation to be applied to the new province. It was to be administered by an Armenian Governor-General appointed by the Porte with the consent of the Powers.

[298] The statistics of population were supplied by the Patriarch Nerses to Mr. Goschen. They may be found in Blue-book, Turkey, No. 23, 1880, p. 274. Mr. Goschen, writing to Lord Granville on 15th July 1880, says: "My strong feeling is that the Powers cannot commit themselves to any plan until they know the real facts about the population. It would not do to build on a mistaken basis, and I feel convinced that no one has sure ground. The Patriarch's figures are as exaggerated as those of the Porte on the other side. Again, how to deal with the nomad Kurds? All must depend on the physical force of the two different races and religions. If the Armenians should be in a minority it will be dangerous to give them the same institutions which we should give if they were in a majority, dangerous to themselves" (Blue-book, Turkey, No. 6, 1881, p. 16). And, again, in another despatch of 23rd July: "With regard to the actual project of reforms, the letter of the Patriarch is conceived in such vague phrases that but little advantage is to be derived from it in elucidating the problem to be solved" (ibid. p. 20).

[299] See Lord Granville's despatch, 10th February 1881: "In consequence of the objections raised by the German Government, Mr. Goschen will not be instructed to put forward the Armenian Question immediately on his return to Constantinople."

[300] The figures for the town and merkez-caza of Van are based on my own knowledge. Those for the other cazas of Van sanjak are the Turkish official figures for 1890, except in the case of Adeljivas caza, where I have substituted a private estimate.

[301] The figures for Bitlis vilayet are the Turkish official figures for 1893.

[302] The figures for Kharput sanjak are an estimate made for me by Consul Boyajean of Diarbekr, at the instance of Consul R. W. Graves. I had previously calculated that the Christians were in a majority in that sanjak. The population of the Dersim sanjak has been estimated from various sources. The estimate is little better than a guess.

[303] The figures for caza Palu have been furnished by Consul Boyajean.

[304] The Turkish official figures, as annexed to the British Consular Trade Report for 1887, have been adopted for the vilayet of Erzerum.

Except in the cases of Van town and caza, and possibly in those of vilayet Kharput and caza Palu, a large percentage might be added to the figures above given in order to provide for the imperfect registration of females. Under this head the figures for the other cazas of Van might be increased by 10 per cent; those for Bitlis vilayet by 13 per cent; and those for Erzerum vilayet by 7 per cent.

[305] The statistical area with which we are dealing for the Turkish provinces measures 42,814 square miles. If we were to adopt the area delimited by the Armenian delegates to the Berlin Congress, the proportion of Christians to Mohammedans would be still smaller.

[306] It is interesting to compare these results, which were obtained quite independently and before I had seen his estimate, with the figures given by the late Mr. Taylor, for many years British Consul for Erzerum and the surrounding country. Mr. Taylor knew the country intimately, and had travelled extensively in it. On his figures are based those which have been given by his successors in office, and which appear in the Blue-books. After making the necessary deductions for districts annexed to Russia since the date of Mr. Taylor's reports, his estimate of the population, as adapted to the area with which we are dealing, is as follows:--Turks, 348,350; Kurds, 466,982; Christians, 352,657--total, 1,167,989. This estimate corresponds in a satisfactory manner with mine, after we have made allowance for information, either new or more complete, which has appeared with reference to certain districts since Taylor's time. The census shows that Taylor under-estimated the Turks who inhabit the northern cazas of Erzerum vilayet. Taylor also placed the Kizilbash Kurds of the Dersim at 110,000. Relying on more recent reports, I place them at 50,000.

[307] Consul Taylor, in alluding to the Kurds of the tableland, has written to the following effect: "The Kurds inhabiting the Erzerum districts, with the exception of the Hakkiari, were originally immigrants from the vicinity of Diarbekr; and there is only one tribe, the Mamakanlu--said to be descended from the Armenian Mamikoneans--who are natives of the soil."

[308] This tax is known in the country under the name of kishlak, or winter quarters.

[309] Vahan Vardapet, in an Armenian newspaper published in Constantinople, the Djeridei Sharkieh, under date the 3/15 December 1886.

[310] See Curzon's Persia, vol. i. p. 548.

[311] Ibid. vol. ii. p. 493.

[312] 2 Kings xix. 37; Moses of Khorene, i. 23.

[313] Saint Martin, Mémoires sur l'Arménie, i. 163.

[314] Faustus of Byzantium, iii. 9.

[315] For instance the Kurdish Beys of Zokh believe themselves to be descended from the dynasty of Sanasar. Again an Armenian convent, called Norshen, is held in reverence by both the Armenian and Kurdish inhabitants; and the name of that convent is believed to be a corruption of Nor-Shirakan or New Shirak--a name applied to the country by the earliest Armenian writers, Agathangelus (ch. cxxvi.) and Faustus (v. 9).

[316] John Katholikos, ch. xxviii.

[317] See especially Turkey, No. I. 1895, parts i. and ii.

[318] It may be found among the archives of the British Consulate at Erzerum.

[319] Chesney, Expedition for the survey of the rivers Euphrates and Tigris carried on by order of the British Government, London, 1850, 2 vols. folio with maps; Narrative of the Euphrates Expedition, London, 1868, 8vo.

[320] "The preservation, so far as it is still possible, of the integrity of Persia must be registered as a cardinal principle of our Imperial creed." "I should regard the concession of a port upon the Persian Gulf to Russia by any power as a deliberate insult to Great Britain, as a wanton rupture of the status quo, and as an intentional provocation to war; and I should impeach the British Minister, who was guilty of acquiescing in such a surrender, as a traitor to his country." "It (i.e. the aggression of Russia upon South Persia and the Persian Gulf) can only be prosecuted in the teeth of international morality, in defiance of civilised opinion, and with the ultimate certainty of a war with this country that would ring from pole to pole."--Curzon, Persia and the Persian Question, London, 1892, vol. ii. pp. 603, 605, 465. May I, as a traveller, take the present opportunity of contributing my mite of gratitude to Lord Curzon for this considerable work?

[321] "The Amir of Afghanistan," Quarterly Review, January 1901, p. 167.

[322] The General Assembly of the Armenian nation met regularly in Constantinople until 1892. Some of the Provincial Assemblies still continue their meetings. But the Constitution is practically in abeyance owing to the strained relations at present existing between the Palace and the Armenians.

[323] Comptes-rendus, Acad. des Sciences, Paris, 1847, xxi. p. 1111.

[324] Vergl. chem. Untersuch. d. Wässer d. casp. Meeres, Urmia u. Van-Sees, Mém. Acad. Sc. St. Pétersbourg, 1859, Séries 6, math. et phys. vol. vii.

[325] Müller-Simonis, P., Du Caucase an Golfe Persique, Paris, 1892, p. 258.

[326] Loc. cit.

[327] Proc. Roy. Soc. lxv. p. 312, London, 1899.

[328] List of Sources.--British Museum Library catalogues; Royal Geographical Society, catalogues and publications; Poole's Index to Periodicals, 1848-96; Review of Reviews, Annual Index, 1890-99; Catalogue of York Gate geographical library, Lond. 1886; Académie des Sciences de St. Pétersbourg, Tableau général des publications (langues étrangères), St. Pet. 1872. List of books in library of Tiflis Museum, kindly compiled for the author by Dr. Radde. Petermann's Mittheilungen, Gotha, 1855-1900; Bibliothèque Asiat. et Afric. (Ternaux-Compans H.), Paris, 1841; Bibliotheca Geographica (1750-1856) (Engelmann, W.), Leipz. 1858; Bibliotheca Orientalis (Zenker, J. T.), vol. ii. Leipz. 1861; Catalogue de la Section des Russicæ, St. Pet. 1873; Bibliographia Caucasica et Transcaucasica (Miansarov, M.), St. Pet. 1874-76; Bibliographies Géog. spéciales (Jackson, J.) Paris, 1881; Orientalische Bibliographie (Müller, A.), 1887-96, Berlin, 1897; Bibliotheca Geographica (Baschin, O.), 1891-97, Berlin, 1899; Catalogue des livres de l'imprimerie arménienne de Saint-Lazare, Venice, 1894.

Encyclopædias: Zedler, Leipz. 1732; Ersch. u. Gruber, Leipz. 1820; M'Clintock and Strong, New York, 1867; Ency. Britannica, Lond. 1875-89; Brockhaus, Leipz. 1882; Meyer, Leipz. 1885; Dictionnaire de Géog. Paris, 1879-95; La Grande Encyclopédie, Paris, 1887 seq.; Real-Ency. f. protestantische Theologie (article Armenien by Gelzer), Leipz. 1897.

Special bibliographies: Dubois de Montpéreux, Voyage autour du Caucase, Paris, 1839-43; Saint-Martin (V. de), Hist. des Découv. Géog. Paris, 1846; Müller-Simonis, Du Caucase au Golfe Persique, Paris, 1892; Bibliographie analytique des ouvrages de M. F. Brosset (1824-79), St. Pet. 1887.

Authorities quoted by Ritter, Die Erdkunde von Asien, Berlin, 1832-59; Reclus, Nouv. Géog. Universelle, Paris, 1876-94; Lanier, L'Asie, Paris, 1889.

End of Project Gutenberg's Armenia (Volume 2 of 2), by H. F. B. Lynch