The oldest testimony we have, apart from Indian folk-lore evidence, to the existence of a collection of stories held together by a framework story like that of our Nights is in the historical encyclopaedia of Mastiidi (d. A.H. 345 = A.D. 956) commonly called "The Golden Meadows." In it (Paris ed. iv., pp. 89f.), speaking of lying stories told by pseudo-traditionalists and by the popular purveyors of edifying legends, he says : "These are like the books transmitted to us and translated for us from the Persian, Indian and Greek, the origin of which was similar to these, such as 'The Book of hazer afsana,' or, translated from Persian to Arabic, 'of a thousand khurcifas,' for a khurcifa is called in Persian an afsdna. The people call this book 'A Thousand Nights and a Night.' " He, then, goes on to give an outline of the framework story which resembles that of our Nights but is not quite the same. Khurcifa is a rather disrespectful Arabic word for a pleasant and strange but incredible story. The exact usage of it has varied at different times. As to the framework story in its different forms, Cosquin (Revue Biblique, January to April, 1909) has shown that a number of similar frames exist in Indian stories. It can be traced definitely back indeed to Indian folklore. Another form of the same passed into North Africa and is the framework story of a collection called The Hundred and One Nights (Gaudefroy Deinombynes Les Cent et Une Nuits, traduites de l'Arabe, Paris, no date). These folk-lore investigations show that any connec tions between this story and the Book of Esther and Persian national legends, according to De Goeje's hypothesis in earlier editions of this encyclopaedia, can only have been en passant when the story was on its way from the extreme east of Asia to the west of North Africa. Cosquin has removed the whole question from literary to folk-lore tradition. Further, it is to be observed that there is no evidence that this Thousand and One Nights of Mas`fidi contained any of the stories which are in our manuscripts of the Nights, except the framework story.
The next witness to a Nights is the Fihrist, a catalogue raisonne of Arabic literature, compiled between A.H. 377 and 400, or, perhaps, slightly later. The author gives a wealth of informa tion as to the origins and development of story-literature in Arabic. He describes the Persian Hazeir afsana with its origin; it contained i,000 nights and less than 200 stories. It existed also in an Arabic translation, but he does not specifically call that "The Thousand and One Nights," as the earlier (Mas`udi) does. He regarded it as a worthless and stupid book, although it is evident that he had not the learned Moslem prejudice against entertaining stories. We may fairly deduce from the evidence that the first Arabic Nights was a straight translation of the Persian Hazeir afsana; that it was a comparatively small book in which each story averaged about five nights ; and that it was not of much value. Again, there is no evidence as to what its stories were except that they must have been of Persian origin.
The next scrap of evidence is derived from an al-Qurti who wrote a history of Egypt under the last Fatimid Caliph al-`Aclid (A.H. 555-567). In it he named specifically a book called A Thousand and One Nights and compared its stories to popular tales current in his time among the people. We have not his
history but we know this passage because it was quoted twice by al-Maqrizi (d. A.H. 845) in his Khitat, i., p. 485; ii., p. 181 and by al-Maqqari (d. A.H. 1041) in his Naf11 i., P. This means that a Nights, of some form or other, was well known in Egypt in Fatimid times.
Next comes the manuscript which Galland used. It was sent to him from Syria after A.D. but had evidently been written in Egypt. From a note in it, it seems that it was at the Syrian Tripoli in A.H. 943, A.D. 1536-37, and it was at Aleppo in A.H. 1001, A.D. 1592-93. When the manuscript was written, before these dates, we do not know. But we can give dates before which it could not have been written, from allusions in certain of the stories in it. In the Cycle of the Porter of Baghdad, at the begin ning of the story of the second calendar, that prince tells that part of his education was in the Sheitibiya, the author of which died A.H. 59o, after al-Qurti's date. On the date of the story of the two Viziers, Nur ad-Din 'Ali and Shams ad-Din Muhammad Prof. William Popper has recently gathered (Journal of Royal Asiatic Society, Jan. 1926) a mass of valuable material. He con cludes that this story cannot be earlier than the reign of Baibars, A.H. 650-76 (A.D. 1260-77), and he is inclined to a date later than A.H. 706. On historical references he would also date the Hunch back Cycle after A.H. 819. Further, in that Cycle, when the Barber makes his apologia in reply to the story told by the young man it is plain that he is speaking after the capture of Baghdad by Hillaga in A.H. 656 (A.D. 1258). But, further again, in the story told by the Christian broker, a Cairene Copt, about his dealings with the young man of Baghdad, the Khan of al-Jawali is mentioned. This Jawali died A.H. 745 (A.D. 1344-45), and the date of the story can hardly be pushed back before that date and may be considerably after that date. See details on all these points in the present writer's article "Earlier History of the Arabian Nights" in Journal of the Royal Asiatic Society pp. 353– 97 (July From the above it is plain that the Nights to which al-Quill testifies as being current in Fatimid times cannot have been the Nights of the Galland manuscript. But the Egyptian recension which Zotenberg has shown was compiled in the later i8th century and which is the present Vulgate of the Nights, so far as editions and translations are concerned, was based demonstrably on a sister manuscript to that of Galland but one more complete as to the number of Nights; the Galland manuscript reaches only Night 281. On the other hand, this modern Egyptian recension has lost grievously in the course of transmission in freshness of vocabu lary and wealth of details. We are left then with the apparent result that some time after A.D. 1400 an Egyptian lover of stories took the framework story of the Nights and built it up into a thousand and one nights, inserting the best stories current in his time that he could find. An incomplete manuscript derived from his recension wandered up into Syria; others more complete re mained in Egypt and formed the basis for the 18th century re cension which Zotenberg identified. Still undefined is the relation ship of these to certain old Syrian manuscripts preserved at Tubingen and Paris, and in the Rylands library at Manchester.