iv. 14), I do not hesitate to identify the site of the temples of

Astishat--Mount Karke, in face of the great range, Taurus--with the immediate surroundings of the present cloister of Surb Karapet (see Vol. II. p. 177). The view which I shall offer from the terrace of that famous monastery (Fig. 157) will be the view which excited the cupidity of the eunuch Hair; the ash trees in the foreground may be the descendants of the hatzeatz-drakht or garden of ash trees; finally, the confluence of rivers, overgrown with thick forest, to which the eunuch descended and where he met his death, may be represented by the still wooded banks of the Murad in the valley of Charbahur. The identification of the scene of the events narrated in the text with the present monastery of Surb Karapet may be found in the geography attributed to Vardan in Saint Martin (Mémoires, etc., vol. ii. p. 431).

The baptism of Tiridates probably took place on the banks of the Upper Murad upon the site of another existing cloister of Surb Karapet, also called Uch Kilisa, near Diadin (see Smith and Dwight, Missionary Researches, p. 417).

[221] Faustus, iii. 3.

[222] Faustus, iii. 13.

[223] Ibid.

[224] Agathangelus, Life of St Gregory, sec. 158.

[225] Faustus, iii. 13.

[226] Faustus, v. 31. "The obsequies of the dead were conducted amid loud lamentations, accompanied by trumpets, guitars and harps. Monstrous dances took place, men and women, with bangles on their arms and painted faces, giving themselves up to every kind of abomination." The picture is coloured by malice, but is vivid.

[227] Agathangelus, Life of St. Gregory, sec. 169.

[228] Faustus, iv. 4.

[229] Faustus, iv. 14. It seems plain from this chapter that these domains had been bestowed upon the family of Gregory by Tiridates and his successors.

[230] Faustus, iii. 15 and 19. The profane delinquents were named Pap and Athenogenes, and the makeshift office-bearers Daniel the Syrian, Pharen and Shahak. The two last-named were formally invested with the office and sent to Cæsarea to be consecrated.

[231] Note especially the election of Nerses I., a descendant of St. Gregory who was loth to accept the office. "The numerous troops of all Armenia demanded and proclaimed Nerses as katholikos ... the entire assembly commenced to cry with a loud voice, 'It is Nerses who must be our pastor.' Nerses refused to accept the mandate, of which he professed himself unworthy. Nevertheless the assembly persisted in their resolution and continued to cry before the king, 'No one except Nerses shall be our pastor; nobody but he shall occupy the holy chair!'" The whole chapter (Faustus, iv. 3) is well worth reading, and contains some very vivid portraiture. Nerses was a layman and was raised to the pontificate in one day. He was then sent to Cæsarea to be formally consecrated.

[232] Professor Gelzer pertinently observes (Die Anfänge, etc., p. 148) that the Armenian kings in pagan times had been in the practice of placing their own near relatives in priestly offices, and quotes Strabo to the effect that in the neighbouring provinces of Cappadocia and Pontus the high priest was deuteros kata timên meta basilea, or second in rank after the king. On the other hand, traces of Jewish custom are to be found in the existence of a second priestly House in Armenia during the early period of Christianity, who in a sense were rivals of the House of the Illuminator. I allude to the House of Albianus. It must not be forgotten that there were extensive settlements of Jews in Armenia at this period, brought thither by the Armenian king Tigranes (Faustus, iv. 55).

[233] Eusebius, Hist. Eccl. ix. 8. For the date see Von Gutschmid (Kleine Schriften, iii. p. 412).

[234] The doubts of Von Gutschmid would perhaps have been removed by the more correct translation given by Professor Gelzer of the passage relating to the journey in Agathangelus and by his editing of the context. The passage should read, "By land and sea they proceeded with haste until they reached the State of the Italians and the land of the Dalmatians and arrived in the imperial city of the Romans." Dalmatia is the præfectura per Illyricum. The name of the bishop is given in the text of Langlois as Sylvester, and as Eusebius in the Greek translation. The best Armenian MS. also has Eusebius. The name of Sylvester appears to have been substituted much later, when the "imperial city of the Romans" was very naturally identified with Rome and the prelate with the bishop of Rome.

My friend Mr. F. C. Conybeare calls my attention to the interesting circumstance that the Armenian equivalent for Latin is Dalmatian. Thus in their Gospels it is said that the title King of the Jews was inscribed on the cross in Hebrew, Greek, and in Dalmatian.

[235] And yet the homoousion was not incorporated into the Armenian Creed! But it does not appear that this omission was intentional. The creed already in use was allowed to stand. I confess to a feeling of astonishment, having regard to the unequivocal language in which the author of the Life attests the acceptance of the Council; but the canons could not have been much appreciated in Armenia at the time if we are to credit Koriun's statement that he himself, with Ghevond and Eznik, brought authentic copies of them to Armenia in the fifth century (Biography of Mesrop in Langlois' Collection, vol. ii. p. 12). Mr. F. C. Conybeare informs me that the Creed of Nice was only communicated to the Armenian diaspora in Persia and Southern Mesopotamia by the Katholikos Papken after the Council of Dvin, c. 490 A.D. It was rejected by that diaspora as in contradiction with their already established Ebionite or Adoptionist tenets (see Letter-book of the Patriarchs, MS. of the Armenian Father, St. Anthony, in Stambul).

Dr. Arshak Ter-Mikelean prints the Armenian and Nicene creeds side by side and accompanies them with some interesting remarks (Die armenische Kirche in ihren Beziehungen zur byzantinischen, Leipzig, 1892, p. 22 seq.). The statement of Agathangelus (Life of St. Gregory), that King Tiridates acted in concert with St. Gregory in making certain additions to the canons must be received with caution, although such additions do appear to have been actually made (see the note of the Mekhitarists to the Italian translation of Agathangelus, p. 196). His son, Chosroes II., appears to have come to the throne in 314. As neither Agathangelus nor Faustus gives us dates, and as the most monstrous anachronisms occur in both treatises, one may do pretty well what one likes with the chronology. I should even mistrust them when they assign a given number of years for a particular period. In the East at the present day ten years means more than one and less than twenty years; and I see no reason to credit the old historians with much greater precision of statement. That the Armenians took part in the Council of Nice is attested by Agathangelus, Faustus, Moses of Khorene, etc., and also by the list of signatures of participants in the Council:--Armeniæ majoris Aristarces, Threnius Diosponti (Von Gutschmid, Kleine Schriften, iii. 415). But we may reasonably doubt that either Tiridates or St. Gregory was alive at the time.

[236] Lazar Pharpetzi, chs. x. and xi.; Moses of Khorene, iii. 36.

[237] Moses of Khorene, iii. 10. The following chronology (which is not that of Moses) is taken from Saint Martin (apud Lebeau, Hist. du Bas-Empire). I attach to it a parallel list of the contemporary Greek Emperors and a similar column for the Sasanian monarchs, which is proudly filled by a single name. The date of Sapor II. rests on the authority of Th. Nöldeke.

Armenian Arsakid Kings.

Chosroes II. (the Little) 314-322 Tiran 322-337 Arshak 338-367 Pap 369-374

Roman Emperors.

Death of Constantine 337 Constantius 337-361 Julian 361-363 Valens 364-378 Theodosius (the Great) 379-395

Persian Sasanian Kings.

Shapur II. (succeeds as an infant) 310-379

[238] Faustus wrote c. A.D. 395-416.

[239] Moses of Khorene (iii. 10) places the king at the head of a Greek army. The patriotism of Faustus was stronger than his veracity, and he maintains a discreet silence upon this circumstance.

[240] The first statement in this sentence is all that we learn from Faustus; the two last rest on the authority of Moses of Khorene, who assigns the death of Verthanes to the third year of Tiran. Aristakes, the younger son of St. Gregory, and his successor in the functions of the pontifical office during the closing years of the life of the saint, was assassinated, apparently by a Roman prefect, at an uncertain date.

[241] In A.D. 339-340, according to Th. Nöldeke (article Persia: Sasanians, in Ency. Brit.).

[242] The peace of A.D. 363.

[243] Agathangelus, Life of St. Gregory, sec. 154.

[244] Faustus, iv. 3.

[245] Mr. F. C. Conybeare has kindly communicated to me the following interesting note to this passage:--"These communities were really cities of refuge, imitated from the old Jewish legislation; and the Armenian monarch's aim was a wise one, namely, to set limits to the blood-feuds and vendettas of his subjects."

[246] I adopt the ingenious suggestion of Professor Gelzer (Die Anfänge, etc., p. 155) that the dioceses of Korduk and Aghdznik were included in the provinces ceded to Persia under Jovian's treaty in 363. Their bishops would have taken refuge in the dominions of the king and be receiving his support. The sequence of events in Faustus is against this hypothesis; but that is not of much account.

[247] We know from Ammianus Marcellinus (xxx. 1) that King Pap himself died in 374.

[248] Professor Gelzer, whose admirable essay I have freely used in the composition of this paragraph, adduces evidence from the correspondence of Basil to show that the advisers of King Pap proceeded cautiously along the path which they had chosen.

[249] Such is the translation given by Professor Gelzer of the passage in Faustus iv. 14.

[250] I am indebted to Mr. F. C. Conybeare for the following note to this passage:--The Armenian alphabet was imposed on Sahak (Isaac the Great) by the Persian Government as a political device to estrange the Armenians both from Greeks and from Syrians. The only historical account is that of Anania of Shirak (unedited chronicle in an uncial MS. at Mush), who relates that the twenty-nine consonants were "arranged in order" by Daniel, a Syrian philosopher, and sent (during the reign of Theodosius the Less) to the Armenian Satrap Vakortsh by Viram Shapu the king by hand of the Elder Abel. The seven vowels were still wanted, and Mesrop received these from Hayek, a noble of Taron. Stephanus, a scribe of Samosata, incorporated these seven vowels among the consonants.

[251] Nor at the Councils of Constantinople and of Ephesus.

[252] It appears that this formula was added to the Trisagion by the Synod of Vagharshapat (Ter-Mikelean, Die armenische Kirche, etc., p. 47).

[253] The subject is fully discussed by Ter-Mikelean (op. cit. pp. 52 seq., and cp. pp. 70 and 89).

[254] My reader may consider that I have been dealing too largely in ancient history. My excuse is that the position remains much the same at the present day. The differences between the Armenian and the Greek Churches are well summarised in a note by the Mekhitarists to the famous address delivered by Nerses of Lambron in the twelfth century to the council assembled at Romkla (Orazione sinodale di S. Nierses Lampronense, Venice, 1812, p. 188). The Greek Church demanded that the Armenian Church should:--1. Anathematise all those who assert that Christ has one nature. 2. Confess Jesus Christ in two natures. 3. Not address the Trisagion to the Second Person of the Trinity. 4. Celebrate the Dominical feasts in conformity with the Greek Church. 5. Prepare the Chrism or Holy Oil with oil alone. 6. Celebrate the Holy Communion with leavened bread and with water in the wine. 8. Receive the fourth, fifth, sixth, and seventh OEcumenical Councils. 9. Receive the nomination of the Armenian patriarch from the Greek Emperor. The attitude of the two Churches towards one another is regretfully but most pithily summed up by the same Nerses of Lambron. The Greeks thanked God that they were not like the Armenians; and the Armenians thanked God that they were not like the Greeks.

It has been generally supposed that the Armenians subscribed the Councils of Constantinople and Ephesus; but I must repeat that this does not appear to have actually been the case (see Ter-Mikelean, op. cit.).

Apart from dogma and ritual, the traveller notices a conspicuous difference between the Greek and the Armenian Church at the present day. You will not find eikons in Armenian houses, while no Russian house is without them. As regards the Church of Rome, the dogmatic breach is even wider than with the Greek Church; in common with the latter the Armenian Church rejects the Filioque. And of course it denies the infallibility of the Pope.

[255] See especially Tournefort, Voyage du Levant, Paris, 1717, vol. ii. p. 335; Parrot, Reise zum Ararat, p. 83, and passim.

The ingenious botanist, Tournefort, was tickled by the question--suggested by the tobacco fields through which he passed--whether the fragrant weed was included among the plants of the terrestrial paradise. Owing to the absence of olive trees in this region, he is puzzled by the story of the dove and the olive branch.

[256] For the intercourse of the Armenians with the Jews I would refer my reader to Ritter, Erdkunde, vol. x. pp. 586 seq.

[257] Dubois, Voyage autour du Caucase, vol. iii. p. 448.

[258] Ibid. p. 419; and compare the account of this city given by Moses of Khorene.

[259] See Ouseley's Travels, vol. iii. p. 450; Ker Porter's Travels, vol. ii. p. 640; Dubois, op. cit. vol. iii. p. 446. Ouseley and Ker Porter thought that they were the remains of Armavir. Dubois probably goes astray in assigning them to Tigranocerta.

[260] Dubois, op. cit. vol. iii. p. 435 seq. On a hill at the confluence of the Arpa Chai with the Araxes, and on the western side of the former river, this traveller found relics of the ancient fortress of Ervandakert. It communicated with the Araxes by a subterranean passage. Ervandashat was situated on the eastern bank, a little higher up the stream.

[261] At Ervandakert and at Karakala, according to the testimony of Dubois. See also Ker Porter (loc. cit.) for the relics of the bridge at the latter place.

[262] Dubois, op. cit. vol. iii. pp. 421 and 449. Compare also Von Behagel's account (apud Parrot, op. cit.).

[263] Probably Sembat II. (A.D. 977-89), the monarch who laid the foundations of the cathedral at Ani.

[264] Ker Porter, op. cit. vol. i. p. 178; Ritter, Erdkunde, vol. x. p. 449.

[265] John Katholikos. He has been translated by Saint-Martin (Paris, 1841, a posthumous work). His History, which for a large part is a history of his own times, does not quite bring us down to the constitution of Ani into a royal residence.

[266] The vanity of the Byzantine court denied them the actual title of king; and the imperial author, Constantine Porphyrogenitus, translates the Persian distinction which they afterwards acquired, that of Shahanshah, by the term archôn archontôn.

[267] For the Artsruni and the Bagratuni I will refer my reader to Saint-Martin (Mémoires sur l'Arménie, Paris, 1818, vol. i. pp. 418 seq.); for the Georgian Bagratuni to Brosset (Histoire de la Géorgie, Histoire ancienne, St. Petersburg, 1849, Addition IX.).

[268] Dulaurier (Recherches sur la Chronologie Arménienne, Paris, 1859, pp. 227 seq.).

[269] Sparapet. This and the other Armenian titles of the age had come down from Arsakid times, having survived the destruction of monarchy. A family retained its title even when the functions which it designated were no longer capable of fulfilment (Saint-Martin, Mémoires, vol. i. p. 420).

[270] The dates which I have taken from Chamchean's History of Armenia I have labelled C. Some are taken from the original work in Armenian; others from the abridged edition translated into English and entitled History of Armenia by Father Michael Chamich, translated by J. Avdall, Calcutta, 1827, 2 vols. 8vo. Those marked D. have been fixed by Dulaurier (op. cit.). Saint-Martin is my authority for some dates.

[271] Thomas Artsruni specifies the length of the various stages in the career of Ashot. See Dulaurier (op. cit. pp. 266 seq.). The date 861 corresponds with the last year of the caliph Mutawakil and the first of the reign of Muntasir. Lane Poole, Mohammedan Dynasties, London, 1894.

[272] Kirakos, quoted by Dulaurier (op. cit.).

[273] For discussions of the site of Bagaran (Pakaran) see Ritter (Erdkunde, vol. x. p. 449), and also Abich (Aus kaukasischen Ländern, Vienna, 1896, p. 203).

[274] Chamchean and Saint-Martin place the death of Ashot in A.D. 889. But see Dulaurier (op. cit. p. 365).

[275] The Tahirids became practically independent in Khorasan A.D. 820-872; they were dispossessed by the Saffarids of Fars and Seistan, A.D. 867-903.

[276] Azerbaijan is, of course, the frontier province of Persia on the side of Armenia, having for capital the city of Tabriz.

[277] Saint-Martin, following Chamchean, attributes another motive to this embassy. Sembat was desirous of severing his connection with the governor of Azerbaijan and of dealing directly with the caliph. Saint-Martin adds that the Caliph Muktafi, who had just succeeded (A.D. 902), granted the request.

[278] Eugène Boré (Correspondance et Mémoires, Paris, 1840, vol. ii. p. 28). The place is situated in the neighbourhood of the town of Erzinjan, and the historian mentions the adjacent village of Tortan, which still appears to exist and to be known under that name. I have not been able to trace it upon any map; but the monastery of Surb Lusavorich and Mount Sepuh, the modern Kohanam Dagh, will be found indicated upon my map, accompanying this work.

[279] Chamchean accounts for this change of policy towards the legitimate king by supposing that Yusuf wished to conciliate him prior to revolting from the caliph.

[280] I adopt the colouring of John Katholikos. Among the many opprobrious terms under which he alludes to Yusuf are the following: second Pharaoh, prince of wild beasts, man-eater, astute serpent, Satan, foul-breathed basilisk. Such is the language of clerical writers in every age.

[281] John Katholikos, ch. clxxxv.

[282] Ibid. ch. clxxxvii.

[283] Samuel of Ani, in Migne, Patrologiæ cursus completus, series Græca, vol. xix. p. 718.

[284] Matthew of Edessa, translated by Dulaurier (Paris, 1858).

[285] Samuel of Ani ap. Migne, op. cit. vol. xix. p. 718.

[286] Matthew of Edessa (op. cit. iii. p. 2) gives the date as A.D. 959-960. He makes the event contemporary with the expedition of the imperial forces against Crete, which started in 960 and was continued during 961. Saint-Martin (op. cit. vol. i. p. 364) assigns the Armenian victory to the latter year, and Chamchean to the year 962.

[287] Matthew of Edessa, op. cit. pp. 14 seq.

[288] Vardan. See Brosset, Ruines d'Ani, St. Petersburg, 1860, p. 102.

[289] Samuel of Ani ap. Migne, op. cit. p. 721.

[290] Ibid.

[291] Matthew of Edessa, chs. xxii. and xxiii.; and Asoghik, iii. 38, quoted by Dulaurier.

[292] Samuel of Ani ap. Migne, op. cit. p. 723.

[293] Samuel of Ani (ibid.) and Asoghik.

[294] Samuel of Ani (ibid.).

[295] Samuel of Ani (ibid. p. 720) and Chamchean. According to Samuel of Ani, it was in A.D. 971 that the patriarch established the seat of his spiritual government at Arghina.

[296] Aristakes of Lastivert (op. cit. ii. pp. 358 seq.) and Matthew of Edessa (op. cit. viii. p. 6).

[297] Matthew of Edessa, op. cit. x. p. 8.

[298] Matthew of Edessa and Aristakes of Lastivert.

[299] When Senekerim of Van ceded his kingdom in A.D. 1021 it had been harried for twenty-two years. Such is the statement of Samuel of Ani (op. cit. p. 723). It is true he attributes these incursions to the "Saracens"; but he must mean the Turks, unless we are to discredit altogether the detailed statement of Matthew of Edessa (ch. xxxviii.), that it was a horde of Turks that defeated the forces of Senekerim. I shall not attempt to reconcile the Armenian accounts with the information which we have received from other sources concerning the early incursions of the Seljuks. The Byzantine writers do not appear to mention the invasions of 1021 and preceding years, or the invasion of 1042 (Brosset ap. Lebeau, Hist. du Bas Empire, vol. xiv. p. 353).

[300] Matthew of Edessa and Aristakes of Lastivert.

[301] Samuel of Ani, Thomas Artsruni (quoted by Dulaurier, Recherches sur la Chronologie Arménienne, pp. 282 seq.), and Chamchean. I prefer to translate oppida by villages and urbes by towns in the Latin version of Samuel of Ani, feeling sure that these terms, as understood in modern times, will be more in accordance with the facts.

[302] Vardan (quoted by Dulaurier, notes to Matthew of Edessa, op. cit. p. 378), and Matthew of Edessa, ch. xi. If Toghrul Bey was over seventy years old when he died in A.H. 455, he would be in the flower of his age at the time of this expedition.

[303] Matthew of Edessa, ch. lx. p. 71; and Chamchean, vol. ii. pp. 127 seq.

[304] Matthew of Edessa, ch. lxix. p. 80. See also Lebeau, op. cit. vol. xiv. p. 351.

[305] The campaigns of this period are narrated by Matthew of Edessa (ch. lxxiii. pp. 83 seq.) and Aristakes (op. cit. pp. 268-82 and p. 285), as well as by the Greek and Arab historians. The subject is discussed by Saint-Martin (Mémoires, vol. ii. pp. 201 seq.).

[306] Matthew of Edessa, ch. lxxviii. pp. 98 seq., and Aristakes, op. cit. 1863, ch. xvi. p. 289. Melazkert owed its deliverance largely to the intrepidity of a Frankish adventurer. It did not fall to the Turks until A.D. 1069, when it was taken after a siege of a single day by Alp Arslan (Matthew of Edessa, ch. cii.).

[307] Matthew of Edessa, ch. lxxxi. p. 109.

[308] Ibid. pp. 107, 108, and Aristakes, op. cit. 1864, ch. xxi.

[309] Matthew of Edessa, ch. lxxxiv. pp. 111 seq.

[310] See Aristakes, ch. xviii., and Matthew of Edessa, ch. lxxxvi.

[311] We are informed in the History of Thomas Artsruni that Senekerim and the Artsrunian princes were accompanied in their emigration by a population of 14,000 males, besides women and children. See Dulaurier, Recherches, etc., p. 284. Chamchean (vol. ii. p. 113) increases this estimate to 400,000 souls, I know not upon what authority.

[312] Chamchean, vol. ii. p. 104; Saint-Martin, Mémoires, vol. i. p. 366; Brosset ap. Lebeau, vol. xiv. pp. 184 seq. Chamchean and Saint-Martin place this expedition in A.D. 999, Lebeau in 991, while Aristakes assigns it to the year 1001. The latter attributes the capture of Arjish to Nikephorus, the Greek governor of Vaspurakan appointed by Basil.

[313] Aristakes in op. cit. ch. ii., together with the authorities collected in the accompanying notes by M. Prudhomme. Chamchean attributes the cession of the kingdom of Ani to the terror which had been inspired by the Seljuk invasions. Basil's policy of taking over the hereditary possessions of the Armenian and Georgian princes and giving them seats in other parts of the Empire was continued by his brother Constantine. See Aristakes, op. cit., third series, vol. xvi. pp. 51 seq.

[314] Samuel of Ani, op. cit. p. 723; and Lebeau, vol. xiv. p. 249. Aristakes is our authority for a curious story respecting the adventures of this testament (ch. x.).

[315] Samuel of Ani; Matthew of Edessa; Aristakes; Kedrenus. The Byzantine historians omit the campaign of 1041, and maintain silence upon the disagreeable topic of the deception practised upon King Gagik.

[316] Aristakes, ch. xvii.

[317] Matthew of Edessa, chs. lxxxiv. and lxxxv.

[318] Aristakes, ch. xvii.

[319] Matthew of Edessa; Samuel of Ani; Aristakes. The king of Kars gave over his realm to the Empire shortly after the fall of Ani, taking in exchange the fortress of Tsamentav near Amasia in Asia Minor (Matthew of Edessa, ch. lxxxviii.).

[320] Pp. 337 and 362.

[321] Kedrenus calls him ruler of Tibion (= Tivin or Dvin) and parts of Persarmenia about the river Araxes (edit. Bekker, vol. ii. pp. 556 seq.). See Matthew of Edessa (ch. x. with Dulaurier's note, and ch. cii. p. 165) and Aristakes (ch. x.). For the Beni-Cheddad see Saint-Martin (Mémoires, vol. i. p. 433; ii. p. 235) and Brosset (Ruines d'Ani, pp. 114 and 126, and Hist. de la Géorgie, Hist. ancienne, p. 343). Abulsevar marched with Alp Arslan in 1069 against the Empire (Matthew of Edessa, ch. cii.). His activity therefore ranges over a considerable period.

[322] Samuel of Ani.

[323] Samuel of Ani; Matthew of Edessa; the Georgian annalist, quoted by Brosset (Hist. de la Géorgie, p. 369).

[324] Samuel of Ani and Matthew of Edessa.

[325] Samuel of Ani.

[326] Samuel of Ani and Matthew of Edessa.

[327] Samuel of Ani; the continuation of Matthew of Edessa; the Georgian annalist in Brosset (Hist. de la Géorgie).

[328] Brosset, Ruines d'Ani, p. 131, and Voyage Archéologique, livraison 1, rapport 1, p. 94.

[329] The Georgian annalist, ap. Brosset, Hist. de la Géorgie.

[330] The various emigrations of the inhabitants of Ani are exhibited by Minas Bejeshkean (Travels in Lehastan (Poland) and other Countries inhabited by Armenian Emigrants from Ani, Venice, 1830 (in Armenian)). His account is summarised by Brosset (Ruines d'Ani, pp. 138 seq.) and by Ritter (Erdkunde, vol. x. pp. 597 seq.). For the code of the Armenians of Lemberg see Sitzungsberichte der phil.-hist. Klasse der k. Akad. der Wissenschaften, Vienna, 1862, pp. 255 seq.

[331] Let me catalogue in this place the works of previous travellers having reference to Ani which I have collected. I shall annex the date of visit whenever I have been able to ascertain it. I have purposely omitted works written in Russian or in Armenian. The full titles will be found in the bibliography attached to Vol. II.

(1) 1621, Poser (Reyse, etc., Jena, 1675, 4o). His account is confined to a few sentences. He mentions the existence of 200 churches in Ani and the immediate neighbourhood. (2) Tavernier (edit. Paris, 1679, Livre Premier, p. 24). A few misleading sentences. (3) 1817, Ker Porter (Travels in Georgia, etc., London, 1821-22, vol. i. pp. 169 seq.). A fantastic description. (4) 1836, Hamilton (Researches in Asia Minor, etc., London, 1842, vol. i. pp. 197 seq.). The best of these older notices. (5) 1837, Wilbraham (Travels, etc., London, 1839, pp. 287 seq.). The hasty but vivid impressions of a tourist, from which the following is an extract: "The shapeless mounds of Babylon are like the skeleton; but the deserted, yet still standing city (Ani) resembles the corpse whose breath has fled, but which still retains the semblance of life." (6) 1837, Abbott (Notes of a Tour, Journal R.G.S., 1842, vol. xii. pp. 215 seq.). Not important. (7) 1838, Eugène Boré (Corr. et Mém., Paris, 1840, vol. ii. p. 2) mentions a mémoire in which he was about to resume the results of his seven days' sojourn in Ani, during which he copied inscriptions. The mémoire has been lost. (8) 1839, Texier (Description de l'Arménie, etc., Paris, 1842, folio, pp. 93-116), with a plan, which is not oriented, and ten fine plates. Texier's account is both defective and unsatisfactory; but it is the first detailed description. I must warn my reader against accepting his history; he seems to confuse Timur with Alp Arslan in some places. (9) 1844, Herrmann Abich (Bull. hist.-phil. de l'Acad. de Sciences de St. Pétersbourg, 1845, vol. ii. pp. 369-76, with notice by Brosset; Aus kaukasischen Ländern, Reisebriefe, Vienna, 1896, pp. 176-200). The distinguished geologist devoted four days to the study of the ruins and drew out a plan of the site. His full account, for which consult the latter of the two references, had not been published, so far as I could ascertain, at the time of my own journey. But Brosset had already published the plan, the substantial accuracy of which I was able to test upon the spot (Voyage Archéologique, St. Petersburg, 1849-51, Atlas), and the inscriptions copied by Abich (in the same work, livr. 1, rapp. 3, pp. 86-111). (10) 1846, Muravieff, quoted by Khanikoff ap. Brosset (Voy. Arch. livr. 1, rapp. 3, pp. 121-52). (11) 1847, Nerses Sargisean of the Society of the Mekhitarists of Venice copied a number of the inscriptions. See Brosset (Ruines d'Ani, St. Petersburg, 1860, p. 5), and especially Brosset's article in the Bull. Acad. Sciences St. P., 1862, vol. iv. pp. 255-67. (12) 1848, Khanikoff copied the Mussulman inscriptions. See Brosset (Voy. Arch. livr. 1, rapp. 3, pp. 121-52). (13) 1850, Kästner (Lieut. Julius) was commissioned by Prince Vorontsoff, Governor of the Caucasus, to explore Ani, and spent forty-four days within its walls. He collected fifty inscriptions and made numerous drawings, which have been made use of by Brosset (Ruines d'Ani, pp. 4 seq.). (14) 18--, Ussher (Journey from London to Persepolis, London, 1865, pp. 243-45). A sketchy description.

The whole subject has been fully treated, but unfortunately at second hand, by Brosset (Ruines d'Ani, St. Pet. 1860, and Bull. Acad. Sciences St. P., 1862, vol. iv. pp. 255-67). The traveller is deeply indebted to Brosset for these two valuable treatises. Fergusson has devoted a few pages to Ani in the first volume of his History of Architecture (see pp. 473-75).

I ought not to close this list without referring to two works in Armenian which are of special value: Sargis Dgalaleantz (Journey in Great Armenia, Tiflis, 1842 and 1858, 8vo), and Alishan (Description of Great Armenia, Venice, 1855). Both these works contain accounts of Ani.

[332] This ravine is the Armenian Tsaghkotzadzor or Valley of the Flower-garden.

[333] The moat may have united the waters of the Alaja and the Arpa Chai. See Ruines d'Ani, p. 60.

[334] See Brosset, Ruines d'Ani, pp. 18 and 144. It may belong to the Tartar period (Mongol) and have reference to the restoration of Ani after the earthquake of A.D. 1319. Texier (op. cit. p. 94) commits himself to the statement that it is in Arabic characters; but see Khanikoff, op. cit. p. 135.

[335] On the authority of Samuel of Ani. See supra, p. 354.

[336] See Brosset, Ruines d'Ani, pp. 16, 17, 58, 59; and Voyage Arch., livr. 1, rapp. 3, p. 143. One of these inscriptions indicates that the name of the reigning prince of the Beni-Cheddad in A.D. 1160, just before the Georgian conquest, was Phatl (Fathlun). Several belong to the reign of Thamar, and exhibit the name of the Georgian ruler, Zakare-Shahanshah, who is styled "chief of the mandatories" and son of Sarkis Shahanshah. See Brosset (Voyage Arch., livr. 1, rapp. 1, pp. 92-94, and Ruines d'Ani, p. 18) for an explanation of this title. Two of these inscriptions of Zakare belong to the years 1206 and 1215 respectively.

[337] Ani is said to have contained not less than 100,000 inhabitants in the eleventh century. Yet the circumference of the city has been estimated at not more than 3 1/2 miles. I am inclined to think that a large proportion of the population lived without the walls.

[338] The conjecture which Brosset throws out that the mosque referred to may be the cathedral is not, I think, a happy one. For this minaret see especially Khanikoff (op. cit. pp. 135-36), Brosset (Ruines d'Ani, p. 31), and Abich (Aus kauk. Länd. vol. i. p. 191). The inscription describes Kei-Sultan as "son of Mahmud, son of Chawir, son of Manuchar, Cheddadi." Kei-Sultan is not otherwise known. We must conclude that the Beni-Cheddad were still powerful in Ani as late as the end of the twelfth century.

[339] The dimensions of the interior are as follows, according to my measurements:--Length, 105 feet 6 in. (viz. 76 feet 6 in. to the daïs of the apse, and 29 feet from the daïs to the extremity of the recess); breadth, 65 feet 6 in.; breadth of apse, 29 feet 7 in.

[340] Texier reminds us that at the time when this cathedral was built (early eleventh century) the Romanesque style was universal in Europe (op. cit. p. 112). Yet in this building we have the characteristics of a style which might be found in Southern Europe in the thirteenth century--the pointed arch, the coupled piers. See also Fergusson, op. cit. p. 473.

[341] I must caution my reader against the drawing of this apse in plate ix. of Brosset's Atlas to the Ruines d'Ani.

[342] The cathedral has been recently constituted into quite a little museum, all fragments of sculptured stone found at Ani being preserved there. I photographed one of the most remarkable, which displays the familiar subject of the eagle and the hare (Fig. 75). Another contains a bas-relief of three saints and was probably placed above a doorway.

[343] Brosset, Voyage Arch. livr. 1, rapp. 3, pp. 93-95, and Ruines d'Ani, pp. 22-28.

[344] Siunik was one of the large provinces into which Armenia was divided. Samuel of Ani places the completion of the cathedral in Arm. era 457 = A.D. 1008. But he may refer to a stage which was not quite the ultimate one.

[345] Brosset identifies this Aron with the Aron-Vestes of the Byzantines, who was sent to these countries about the year 1042, was commander of the imperial forces, became governor of Vaspurakan, Ani being attached to his jurisdiction, and was still in possession of his office in 1048 (Voyage Arch. loc. cit. p. 93).

[346] I am not aware that any inscription mentions the name of the architect. Sic vos non vobis! But Asoghik tells us that it was Tirdat or Tiridates, an Armenian architect who is reputed to have restored St. Sophia at Constantinople after its partial destruction by an earthquake.

[347] My measurements of the interior are:--Length, 41 feet (of which 15 feet is occupied by the apse measured from the daïs to the extremity of the recess); breadth, 26 feet. Texier mentions an adjacent baptistery (?).

[348] See especially Texier, Muravieff, and Abich's Aus kauk. Länd. vol. i. pp. 198-99.

[349] The inscription has been translated by Brosset (Ruines d'Ani, pp. 145-48).

[350] Brosset (ibid. p. 15). I was able to verify the date, about which Brosset expresses some doubt (ibid. p. 148).

[351] For these two ruins see also Abich, op. cit. vol. i. pp. 196-97.

[352] For these inscriptions see Brosset, Ruines d'Ani, pp. 11-13. He reminds us of the importance of the date 1320 (Arm. era. 769) as being the year after the great earthquake. I must take this opportunity to caution my reader against accepting the tradition mentioned by Muravieff (ap. Brosset, Voyage Arch. livr. 1, rapp. 3, p. 127) that the little chapel was built in A.D. 1000 by King Gagik I. I may also mention that we could discover no traces of the guardhouse adjacent to the bridge (Ruines d'Ani, p. 10).

[353] Khanikoff ap. Brosset, Voyage Arch. livr. 1, rapp. 3, p. 138.

[354] Ibid. p. 138, and Ruines d'Ani, p. 30.

[355] Ibid. p. 140, and Ruines d'Ani, p. 31.

[356] Muravieff ap. Brosset, Voyage Arch. loc. cit. p. 129.

[357] Mr. Marr has published an account of his discoveries of new epigraphical material in Armenia in the Zapiski of the Eastern Section of the Imp. Russian Arch. Society, vol. viii., 1893, pp. 69-103. He contributes four new inscriptions from Ani. I have not been able to find any account of his excavations.

[358] The interior of the building which forms the subject of my illustration is given by Brosset in plate xiv. of the Atlas to the Ruines d'Ani. The detail and ornament there portrayed do not correspond with reality. The devils are more or less imaginary, and there appears to be only one of them in the actual design, viz. on the south wall, the first pilaster as you enter from the west--in low relief. Brosset styles this interior "a hall in the citadel"; but the following considerations are against this view:--1. It is oriented east; 2. It obviously had an apse; 3. Above the apse you see the form of a cross sculptured on the face of the arch which still remains.

The bas-reliefs are given by Brosset, plates xxxv. and xxxvii. The former (representing the archer) was found in the valley of the Tsaghkotz with an inscription in Armenian, "Christ have pity on the lady Shushan, thy servant." This personage may be identified with the wife of the Pahlavid Grigor, mother of Vahram.

[359] This building must be the subject of plate xiii. in Brosset's Atlas to the Ruines.

[360] The rock with the chapel is described by Abich (op. cit. vol. i. p. 192). It was strongly fortified.

[361] It is not exactly symmetrical, the measurement from west to east being nearly 31 feet.

[362] Brosset translates, "J'ai construit ce lieu de repos." But it surely cannot refer to the chapel itself, which, as we have seen, has inscriptions of the mother of Aplgharib, and must therefore have been in existence before 1040. Brosset therefore supposes that the restoration of the church is alluded to (Ruines d'Ani, pp. 37 and 106). For a more probable version of the inscription see Alishan, Shirak, p. 53.

[363] For the inscriptions see Brosset (Voyage Arch. loc. cit. p. 91, and Ruines d'Ani, pp. 36 seq.). Aplgharib was brother to Vahram. I could find no trace of the curious figure found upon one of the windows which Brosset refers to (pp. 38 seq.). On the other hand, I was able to identify the two inscriptions last mentioned.

[364] Kirakos ap. Dulaurier, Recherches sur la Chron. Arm. p. 280.

[365] Asoghik ap. Brosset, Ruines d'Ani, p. 106.

[366] Abich confuses the sites of these two monuments in his Reisebriefe (op. cit.).

[367] Such is the translation of this inscription given by the editor of Aristakes of Lastivert. Brosset appears to have made a palpable error (Ruines d'Ani, p. 21, inscription of Christaphor).

[368] Probably the inscription of this same Aplgharib given by Brosset (Ruines d'Ani, p. 28) belongs to this chapel. It runs thus:--"Under the pontificate of Ter Petros and the reign of Sembat son of Gagik Shahanshah in the year 485 (A.D. 1036) I, the Marzpan Aplgharib, son of the prince Grigor, grandson of Apughamir and brother of Vahram and Vasak, constructed at great expense this Surb-Phrkich in the metropolis of Ani." This inscription would establish as a fact that the chapel itself was dedicated to the Redeemer.

[369] A perfect labyrinth of confusion has been brought into existence by the attribution of the east front of the portal of the church of the Apostles to this castle or palace (see plate xix. of Brosset's Atlas). Happily I am able to correct the error. It has been instrumental in leading Brosset to assign all the inscriptions found in that church to this castle. The name "palace of the Pahlavids" is purely imaginary.

[370] Brosset, Ruines d'Ani, p. 51, and plate xxxvi. No. 3 of the Atlas. It has been wrongly attributed to the castle.

[371] Abich describes this chapel as "a magnificent church in the form of a Greek cross with a central rotunda and four large semicircular niches at the sides" (op. cit. vol. i. p. 190).

[372] See Brosset (Voyage Arch. livr. 1, rapp. 3, pp. 86, 100, 101, 106, 109; and Ruines d'Ani, pp. 48-52).

[373] This is the chapel which Abich names "Kirche der Maria Verkündigung" (op. cit. vol. i. p. 193).

[374] Abich, op. cit. vol. i. p. 199.

[375] See the Georgian annalist translated by Brosset (Hist. de la Géorgie).

[376] I should be sorry to have to swear to this statement.

[377] Voyage Arch. livr. 1, rapp. 3, pp. 96, 107, 109-10.

[378] Telfer, Crimea and Transcaucasia, vol. i. p. 216. The chamber at Geghard is known as the Rusukna sanctuary, and was completed in A.D. 1288 (Arm. era 737) (ibid.).

[379] An inscription of A.D. 1215, much mutilated, seems to infer this (Brosset, Voyage Arch. loc. cit. p. 97).

[380] Brosset, Voyage Arch. loc. cit. p. 98. The dimensions of these various apartments are:--No. 1, length, 29 feet 4 inches; breadth, 29 feet; No. 2, 27 feet by 27 feet 2 inches; No. 3, hall of the synod, 18 1/2 paces by 18 paces. The reader will note that the architects avoided exact squares. In this they were governed by a right instinct.

[381] Brosset, Voyage Arch. loc. cit. p. 99. Another derivation is from the Greek word for a priest, iereus (see M. Prudhomme, note to Aristakes, ch. ii.).

[382] Asoghik ap. Brosset (Ruines d'Ani, p. 137).

[383] Ruines d'Ani, p. 137.

[384] Ibid. p. 61.

[385] Texier (op. cit. p. 112):--"La façade de cette église (the cathedral) construite avec une simplicité remarquable ... peut être regardée comme le type de l'architecture allemande du moyen âge. Il est facile d'expliquer comment, dans toute cette contrée, on retrouve le dôme à toit conique particulier à l'architecture arménienne. En effet, après la prise d'Ani par les Mussulmans, un grand nombre de citoyens abandonnaient la ville...."

[386] Abich, Geologische Forschungen in den kaukasischen Ländern, Vienna, 1882, part ii. pp. 47 seq. Das Plateau von Kars.

[387] Strabo, xi. c. 528.

[388] Ptolemy, v. 13, pages 135 and 136 of the folio edition. The identification with one of these towns is generally assumed; but in view of the statement of Evliya, noted below, that in his time there existed three towns of this name, it cannot be regarded as certain.

[389] to kastron to Kars, Const. Porphyr. De adm. imp. cap. 44.

[390] See Chapter XVIII. p. 353 and p. 364; and Saint-Martin, Mémoires sur l'Arménie, vol. i. p. iii. Tsamentav was the name of the appanage received in exchange. It was situated in the Cilician Taurus.

[391] Koch, Reise im pontischen Gebirge, Weimar, 1846, p. 462.

[392] Travels of Evliya, translated by Von Hammer, vol. ii. p. 181. The passage runs: "Eight hours further to the east we reach the frontier fortress of the Ottomans, the castle of Karss. There are three towns of that name; one is in Silefka, the Karss of Karatashlik; the second the Karss of Mera'ash, and the last that of Dúdemán, which is the present one." I am ignorant of the locality assigned to the first mentioned.

[393] The name Vanand is said by Moses of Khorene (ii. 6) to be derived from that of the chieftain of a horde of Bulgarians who settled there. Now that Moses has been assigned to the eighth century of our era the statement need not surprise us.

[394] Von Hammer, Geschichte des osm. Reiches, vol. viii. p. 58.

[395] Uschakoff, Geschichte der Feldzüge, 1828, 1829, Leipzig, 1838,