been superseded by the discussion of F. Görres in the Zeitsch. f. wiss. Theologie, xvi. 1890, p. 454 sqq. Ritter St. Georg in Geschichte, Legende, u. Kunst." [There is no question that the Acta (in Act. Sanct. 23rd April) are apocryphal and legendary. They are remarkable for the horrible descriptions of scenes of martyrdom, which might serve as a text to elucidate the pictures on the walls of the curious round Church of San Stefano on the Esquiline.] Görres arrives at practically the same conclusion as Tillemont (Mém. eccl., v. 185-9, 658-60). All the details of St. George's martyrdom are uncertain; but St. George existed and suffered as a martyr in the East in some pre-Constantinian persecution. Tillemont established the reality of St. George by the existence of his cult (he was a Meyaλóμaprus) in the sixth century; Gorres proves that it already existed in the fifth century. (1) The round Church of St. George at Thessalonica is not younger than the fifth century and possibly belongs to the fourth; (2) Venantius (Carm. ii. 12, p. 41, ed. M. H. G.) mentions a Gallic basilica to St. George, founded by Sidonius Apollinaris; (3) the decree of Pope Gelasius de libris non recipiendis, at end of fifth century, condemns the Acta of St. George as apocryphal, but confesses his historical existence.

The connexion of his name with a dragon-slaying legend does not relegate him to the region of myth. For over against the fabulous Christian dragon-slayer, Theodore of the Bithynian Heraclea, we can set Agapetus of Synnada and Arsacius, who though celebrated as dragon-slayers were historical persons.

Gibbon's theory which identifies St. George with George of Cappadocia has nothing to be said for it; but Görres points out that it is not open to any objection on the ground that George of Cappadocia was an Arian. For there are examples of Arians admitted into the Martyrologium: he cites Agapetus of Synnada and Auxentius, afterwards bishop of Mopsuestia. (It is to be noted that one recension of the Acta S. Georgii was edited by Arians.)

23. THE CHURCHES OF CONSTANTINE AT JERUSALEM—(P. 455)

In regard to Constantine's Churches at Jerusalem it may be said, without entering upon the question as to the true positions of Golgotha and the Holy Sepulchre, that it is certain that these Churches-(1) the round Church of the Anastasis which contained the Sepulchre, and the (2) adjacent Basilica, dedicated to the Cross-stood on the site of the present Church of the Holy Sepulchre. Injured by the Persians (614 A.D.) they were restored some years later, and a plan of the buildings drawn up, towards the end of the seventh century, by the pilgrim Arculfus is extant, and is of great importance for the topography. Some traces of the old buildings still remain. "The relative position of the Churches is the same; the circular Church of the Anastasis has preserved its form; the south wall of the Basilica can be traced from 'Calvary' eastward, and one of the large cisterns constructed by Constantine has been discovered" (Sir C. Wilson, in Smith's Dict. of the Bible, new ed., 1893, p. 1654). Mr. Fergusson's theory which identified the Church of the Resurrection with the mosque known as Kubbet-es-Sakhrah, the Dome of the Rock (within the so-called Haram area"), is now quite exploded.

The Dome of the Rock has its own question, but has nothing to do with Constantine. Is it of Saracenic origin dating from the end of the seventh century-built perhaps by a Greek architect? or was it originally a Christian Church, and converted into a mosque? It has been identified by Professor Sepp with a Church of St. Sophia built by Justinian. Sir C. Wilson thinks that it stands on the site of St. Sophia, which was destroyed by the Persians; "that it was rebuilt with the old material by Abdul-Melik who covered it with a dome, and that it was again repaired and redecorated by El Mamûn" (ib., p. 1657).

The adjacent mosque el-Aksa occupies the site of the mosque of Omar. It was built by Abd al Malik, "out of the ruins of Justinian's Church of St. Mary" (Wilson, ib.), which is fully described by Procopius; but there is a difference of opinion, whether the Church was on the same site as the mosque or 36*

VOL. II.

(so Fergusson and others) in the south-eastern corner of the "Haram area," where there are vaults apparently of the Justinianean age.

For further details see Sir C. Wilson's article Jerusalem, cited above; Mr. T. H. Lewis' essay on the Church of Constantine at Jerusalem in the Palestine Pilgrims' Text Society, 1891; Sepp, Die Felsenkuppell eine Justinianische Sophien-kirche; various papers in the Palestine Exploration Fund publications.

24. THE TIGRIS AND EUPHRATES-(P. 495)

The recent publication of a geographical description of Mesopotamia and Baghdad by an Arabic writer, Ibn Serapion, of whom nothing is known except that he wrote in the early years of the tenth century, by Mr. Guy Le Strange (with translation and commentary, in the Journal of the Royal Asiatic Soc., 1895, January and April; cp. addenda in July, and 1896, October), is of considerable importance.

It shows that since the tenth century great alterations have taken place in the course of the Tigris and Euphrates, and shows what these alterations were; it gives a clear account of the canal system which drew the overflow of the Euphrates into the Tigris; and it supplies most important data for the reconstruction of the topography of Baghdad.

Before the Caliphate, the River Tigris followed its present course, from Kut-al-Amarah (about 100 miles below Baghdad) flowing in a south-easterly direction to its junction with the Euphrates. But during the middle ages-in the tenth century for example-it flowed almost due south "running down the channel now known as the Shatt-al-Hay, and passing through the city of Wasit" (Le Strange, ib., Jan., p. 3). The changes in the Euphrates are thus summed up by Mr. Le Strange (p. 4): A little above Al-Küfa "the stream bifurcated. The branch to the right-considered then as the main stream of the Euphrates, but now known as the Hindiyya Canal-ran down past AlKufa, and a short distance below the city became lost in the western part of the great Swamp," which also swallowed up the waters of the Tigris. "The stream to the left or eastward called the Sura Canal-which, in its upper reach, follows the line of the modern Euphrates-ran a short course and then split up into numerous canals whose waters for the most part flowed out into the Tigris above Wasit." The great Swamp in which the streams of both Tigris and Euphrates lost themselves was drained by the Tidal Estuary which reached the sea at Abbadān, "a town which, on account of the recession of the Persian Gulf, now lies nearly twenty miles distant from the present shore-line".

It should be carefully remembered in reading the account of the events after Julian's death that the Tigris has also altered its course to the north of Ctesiphon since the tenth century. From a point below Samarra to a point above Baghdad, it followed a shorter and more westerly channel than at the present day.

As to the canal Nahr-al-Malik (see above, p. 503), Mr. Le Strange says (ib., Jan., p. 75), that "roughly speaking it followed the line of the modern Radhwaniyya Canal”.

It may be added that the geographical work of Abu-l-Fidă, mentioned by Gibbon, p. 495, n. 54, is not very valuable, being neither good nor early. The authoritative Arabic text is that of Reinaud, 1840, and there is a French translation by S. Guyard, 1883. On early geographical works in Arabic, see Le Strange's Palestine under the Moslems (Pal. Explor. Fund).

THE ABERDEEN UNIVERSITY PRESS LIMITED

ADDENDA

P. 304, 1. 17, the luminous trophy of the cross, placed above the meridian sun. [I cannot forbear to mention here the ingenious and plausible suggestion communicated to me by Professor Flinders Petrie that what Constantine saw was the phenomenon of mock-suns (not uncommon in northern, but rare in southern, latitudes). The real sun, with three mock-suns, might have appeared to his eyes as a cross.]