desert (al-Rub‘ al-Khálí, the 'Solitary Quarter') which stretches across the Arabs of the North and South. peninsula and forms a natural barrier to intercourse, but also by the opposition of two kindred races widely differing in their character and way of life. Whilst the inhabitants of the northern province (the Ḥijáz and the great central highland of Najd) were rude nomads sheltering in 'houses of hair,' and ever shifting to and fro in search of pasture for their camels, the people of Yemen or Arabia Felix are first mentioned in history as the inheritors of an ancient civilisation and as the owners of fabulous wealth—spices, gold and precious stones—which ministered to the luxury of King Solomon. The Bedouins of the North spoke Arabic—that is to say, the language of the Pre-islamic poems and of the Koran—whereas the southerners used a dialect called by Muḥammadans 'Ḥimyarite' and a peculiar script of which the examples known to us have been discovered and deciphered in comparatively recent times. Of these Sabæans—to adopt the designation given to them by Greek and Roman geographers—more will be said presently. The period of their bloom was drawing to a close in the early centuries of our era, and they have faded out of history before 600 a.d., when their northern neighbours first rise into prominence.

It was, no doubt, the consciousness of this racial distinction Ishmaelites and Yoqṭánids. that caused the view to prevail among Moslem genealogists that the Arabs followed two separate lines of descent from their common ancestor, Sám b. Núḥ (Shem, the son of Noah). As regards those of the North, their derivation from ‘Adnán, a descendant of Ismá‘íl (Ishmael) was universally recognised; those of the South were traced back to Qaḥṭán, whom most genealogists identified with Yoqṭán (Joktan), the son of ‘Ábir (Eber). Under the Yoqṭánids, who are the elder line, we find, together with the Sabæans and Ḥimyarites, several large and powerful tribes—e.g., Ṭayyi’, Kinda, and Tanúkh—which had settled in North and Central Arabia long before Islam, and were in no respect distinguishable from the Bedouins of Ishmaelite origin. As to ‘Adnán, his exact genealogy is disputed, but all agree that he was of the posterity of Ismá‘íl (Ishmael), the son of Ibráhím (Abraham) by Hájar (Hagar). The story runs that on the birth of Ismá‘íl God commanded Abraham to journey to Mecca with Hagar and her son and to leave them there. They were seen by some Jurhumites, descendants of Yoqṭán, who took pity on them and resolved to settle beside them. Ismá‘íl grew up with the sons of the strangers, learned to shoot the bow, and spoke their tongue. Then he asked of them in marriage, and they married him to one of their women.3 The tables on the opposite page show the principal branches of the younger but by far the more important family of the Arabs which traced its pedigree through ‘Adnán to Ismá‘íl. A dotted line indicates the omission of one or more links in the genealogical chain.4

It is undeniable that these lineages are to some extent fictitious. There was no Pre-islamic science of genealogy, so that the first Muḥammadan investigators had only confused and scanty traditions to work on. They were biassed, moreover, by political, religious, and other considerations.6 Thus their study of the Koran Character of Muḥammadan genealogy. and of Biblical history led to the introduction of the patriarchs who stand at the head of their lists. Nor can we accept the national genealogy beginning with ‘Adnán as entirely historical, though a great deal of it was actually stored in the memories of the Arabs at the time when Islam arose, and is corroborated by the testimony of the Pre-islamic poets.7 On the other hand, the alleged descent of every tribe from an eponymous ancestor is inconsistent with facts established by modern research.8 It is probable that many names represent merely a local or accidental union; and many more, e.g., Ma‘add, seem originally to have denoted large groups or confederations of tribes. The theory of a radical difference between the Northern Arabs and those of the South, corresponding to the fierce hostility which has always divided them since the earliest days of Islam,9 may hold good if we restrict the term 'Yemenite' (Southern) to the civilised Sabæans, Ḥimyarites, &c., who dwelt in Yemen and spoke their own dialect, but can hardly apply to the Arabic-speaking 'Yemenite' Bedouins scattered all over the peninsula. Such criticism, however, does not affect the value of the genealogical documents regarded as an index of the popular mind. From this point of view legend is often superior to fact, and it must be our aim in the following chapters to set forth what the Arabs believed rather than to examine whether or no they were justified in believing it.

'Arabic,' in its widest signification, has two principal dialects:—

1. South Arabic, spoken in Yemen and including Sabæan, Ḥimyarite, Minæan, with the kindred dialects of Mahra and Shiḥr.

2. Arabic proper, spoken in Arabia generally, exclusive of Yemen.

Of the former language, leaving Mahrí, Socotrí, and other living dialects out of account, we possess nothing beyond the numerous inscriptions which have been collected South Arabic. by European travellers and which it will be convenient to discuss in the next chapter, where I shall give a brief sketch of the legendary history of the Sabæans and Ḥimyarites. South Arabic resembles Arabic in its grammatical forms, e.g., the broken plural, the sign of the dual, and the manner of denoting indefiniteness by an affixed m (for which Arabic substitutes n) as well as in its vocabulary; its alphabet, which consists of twenty-nine letters, Sin and Samech being distinguished as in Hebrew, is more nearly akin to the Æthiopic. The Ḥimyarite Empire was overthrown by the Abyssinians in the sixth century after Christ, and by 600 a.d. South Arabic had become a dead language. From this time