The Site numbers used in the present study are marked in the general map accompanying the first article (BSOAS xv, 1953, p. 28). More detailed maps of the central Eastern Desert and the Eastern Delta appear below, p. 212. The E. Delta map shows two new site numbers: (a) Site 17 (at El-Maḥemdīlya = Gerra ?) is the spot where Clédat actually found inscription No. 82 and it may be that Site 16 (at ancient Mons Casios) does not figure in the present study (see classical notes, p. 242). (b) Site 18 (QaṣEsr Gheiṭ), east of El-Qanṭara on long. 33° E., has a Nabatsæn fragment, No. 83. Nos. 50–1 and 52 (at Sites 11 and 9, see BSOAS xv, p. 17) are briefly repeated here for convenience, as. Winkler's inscriptions from the Leucos Limen road came to light after the first article had gone to press.

Arabic documents containing eyewitness reports on historical events are very rare. They are the more welcome when they complement and illustrate the often very deficient literary sources. This is the case with the excerpts from two business letters of the 12th century published here, one sent from Aden to India and one from the same town to Cairo. Both passages describe one and the same event: the attack on Aden by a fleet sent by the King of Kīsh with the aim of taking the town or at least a part of it.

Abū ‘Īsā Muḥammad b. ‘Īsā b. Sawra b. Shaddād b. al-Ḍaḥḥāk al-Sulamī al-Būghī al-Tirmidhī is famous for bis collection of traditions which is one of the six books now generally recognized by Sunnīs. He is reported to have said that his grandfather belonged to Marw, but removed from there in the time of Al-Laith b. Sayyār to Tinnidh, a town on the north bank of the Āmū Daryā. Like many others Tirmidhī travelled in pursuit of learning, and heard traditions from numerous Khurāsānīs, ‘Irāqīs, and Ḥijāzīs. He was a pupil of Bukhārī, and also heard traditions from some of Bukhārī's shaikhs, such as Qutaiba b. Sa‘īd al-Baghlānī (d. 240), ‘Alī b. Ḥujr al-Marwazī (d. 244), Abū Kuraib Muḥammad b. al-‘Alā’ al-Kūfī (161–248), and Muḥammad b. Bashshār al-Baṣrī (167–252).4 An interesting story is told to illustrate his power of committing traditions to memory. On the way to Mecca he met a shaikh from whose traditions he had copied out two juz’.

Before broaching my subject I should like to explain the reason why a serial number has been added to the title of this article. A list of my publications will be found in the BSOAS., 1953, xiv/3, pp. 69–81, but as a variety of subjects has interested me during my career, readers may be baffled by the seemingly disconnected array of titles. In fact, my original intention was to integrate my studies in separate books, but on seeing that for both material and personal reasons this plan had no chances of success, I decided to take every opportunity for publishing the material collected—not only to solve some definite problems, but also to show the methods which I would have wished to apply to my subjects, in the light of my experience and the observations made in some more developed fields. This underlying preoccupation is what to my mind connects my writings. On the other hand, I should like to draw the attention of users of my ‘ specimens ’ to some particular classes of my studies; in a word, to their serial character.

The gerund here studied is a grammatical form of great interest on account of its phonological variations and the wide range of its syntactical relationships. It occurs in the single form, gide, in a strengthened form, giderek, and in a pair form, gide gide. We find it as an independent gerund and in verbal combinations which vary from mere association to complete fusion, such as gidememek and gideyor (gidiyor). It also has a tendency to become specialized in use as a post-gerund, and numerous particles have originated in this way. It has had a long history, being one of the basic grammatical forms of Turkish, a common feature of the Turkish family of languages from the time of the early inscriptions of northern Asia. In Ottoman Turkish its importance (except in the -erek and pair forms) has gradually declined during the last 500 years, and it is now used only to a very limited extent; but in Old Ottoman it was fairly widely employed, and the texts which have survived provide ample material for study. It is, however, a complicated study. In the work which has been done it would seem that investigators have usually relied on insufficient data, or have applied methods unsuitable for this case. For example, the method of general comparison between Ottoman on the one hand and various speech forms such as Uighur, Chaghatai, and modern Anatolian dialects on the other, cannot give satisfactory results. There are too many unknown factors, too many imponderables. Such a procedure inevitably leads to the comparison of mere isolated examples which do not really prove anything. I have therefore confined myself to Old Ottoman, and my method has been to take whole texts and collect and classify all the examples of this gerund which they contain. This is laborious, but it produces results which cannot be obtained in any other way.

Part I concluded with the suggestion that certain general-phonetic features, those of rounding and spreading, and those of vowel closure, might profitably be given a prosodic statement for the word as a whole, as:—

I. a 3-term system: y/w/ә

II. a 2-term system: o/c

A given word might thus be assigned to a prosodic category, to the ‘ y ’, the ‘ w ’, or the ‘ ә ’ term, and treated as a ‘ y word ’, ‘ w word ’, or ‘ neutral word ’ (yW, wW, әW). Similarly a given word might be assigned to either term of the closure system, and treated as a ‘ close word ’ or as an ‘ open word ’ (cW, oW). Theoretically it is a matter of indifference which of the two systems should be stated first. Since, however, they cannot be stated simultaneously in practice, the method of statement adopted here gives priority to the w/y/ә system; the exponents of ‘ w ’, ‘ y ’, and ‘ ә ’ are stated in turn. The o/c system is considered within the framework of these three categories, the exponents of o/c being given for wW, yW, and әW in turn, immediately after the exponents of ‘ w ’, ‘ y ’, and ‘ ә ’ respectively.

When Buddist works in Sanskrit were first introduced into Europe, it was at once obvious that the language of some of them as it appears in the manuscripts was, in comparison with Classical Sanskrit, frequently ungrammatical, and on occasion barbarously so. The immediate and natural reaction of scholars accustomed to the regularity of Sanskrit was to stigmatize these shortcomings, and to attempt to remove as many irregularities as possible by forcibly emending the text. It was, however, very soon recognized that many of the seeming anomalies could not be abolished, and that they must be accepted as genuine in their own context. This was especially clear in the case of the verses of some of the older texts, where the metre often guaranteed non- Pāηinean forms; and the language of these verses, variously called the Gāthādialect, mixed Sanskrit, or hybrid Sanskrit, was recognized as something in its own right. The same courtesy was readily extended to the prose of the Mahāvastu, which in places could only have been made to resemble Sanskrit by completely rewriting the text. The prose of the other texts, being in many waysvirtually Sanskrit, took considerably longer to win the same recognition; but for many years now it has been generally admitted that here also is a language which must be judged according to its own standards, and not exclusively by the canons of classical Sanskrit grammar.

European students of Yoruba are usually slow in learning to understand the spoken language because, apart from having to grapple with the unfamiliar phenomenon of tone, they often find it difficult at first hearing to recognize in medial position words with which they have become quite familiar when occurring initially or finally. This is due to the frequent occurrence of junction situations where a word with vowel final1 is in close syntactic relationship with a following word which, in its isolated form, has a vowel initial, resulting sometimes in the elision of one or other of the vowels in contact, sometimes in an assimilation of one vowel to the other, and sometimes in each vowel retaining its identity. The tones of the junction syllables may also be elided or modified. The European, with his literate approach to language study, is further handicapped by the anomalies of the accepted orthography, which in some cases indicates the junction phenomena and ignores them in others. An added complication is that in many Yoruba books it is hard to find two consecutive pages which are consistent in maintaining the same set of anomalies. This means that, while in some cases the orthography provides a sufficient clue to the pronunciation of a word, in other cases a word cannot be correctly pronounced without reference to the surrounding words and without a detailed knowledge of the usages of the language.

Two articles of mine on the poetry of Ḥāfiẓ1 were recently criticized severely, even violently, by Dr. N. E. M. Boyce in an article written in a kind of informal collaboration with Professor W. B. H. Henning.2 Dr. Boyce was good enough to send me a page-proof copy of her article very shortly before publication, and I noted at the time how widely divergent were our respective views, both as regarded general approach and also in matters of detailed interpretation. This, however, occasioned no surprise, and in the normal way I should probably have been content to bring my observations privately to Dr. Boyce's notice; but, for personal reasons, I eventually decided to circulate a reply on my own initiative, for the information of colleagues who might have had the original critique brought to their notice before its publication.

I realized that, by taking this action, I might prejudice any chance the reply had of appearing subsequently in print. But such prospects already appeared slender in view of my article's length and largely negative value; and the ordinary delays attending publication appeared even more unwelcome than the probability that the reply might never be read at all in printed form.

Ry. 506 is an inscription of some importance for an otherwise obscure period of Arabian history. G. Ryckmans, in his editio princeps of the inscription,3 and J. Ryckmans in his comments thereon,4 have made very valuable contributions to the understanding of this far from easy text. A few notes may, I think, be nevertheless added.

At the outset, one point to be remembered in dealing with this text is that Abraha, its author, is frequently referred to in the plural, as we see from sṭrw in line 2 here (with Abraha as the subject), as well as throughout C. 541. Consequently, plural verbs and the plural pronoun -hmw can refer to Abraha himself.

Line 1. It is interesting to confront the form of titulature here used with that found in C. 541/4–8. The comparison allows us to make some deductions about the name or epithet zybmn (so spelt here, against zbymn in C. 541):—

In a learned and thoughtful book by Professor Gelb, of Chicago (A Study of Writing: the Foundations of Grammatology2), which I am quite incompetent to review as a whole, there are a number of misleading statements in the section headed Japanese Syllabary. The excuse for drawing attention to them here is that if uncorrected they may put readers who are without a knowledge of Japanese onto false tracks.

That the katakana, ‘ side kana’, is ‘so named because it appears chiefly in addition to the Chinese word signs’ (p. 159) is not the generally accepted explanation, nor does it seem very likely. The reason usually given for the name is that the katakana signs are parts, generally one or the other side, of Chinese characters.