Balochki and Pukhtu or Afghan, besides the direct descendants of the Zend, the Pehlevi, Dari and so forth.  Even in the most barbarous jargons he will find terms which throw light upon the literary Iranian of the lexicons:  for instance “Madiyan” = a mare presupposes the existence of “Narayan” = a stallion, and the latter is preserved by the rude patois of the Baloch mountaineers.  This process of general collection would in our day best be effected after the fashion of Professor James A. H. Murray’s “New English Dictionary on Historical Principles.”  It would be compiled by a committee of readers resident in different parts of Persia, communicating with the Royal Asiatic Society (whose moribund remains they might perhaps quicken) and acting in co-operation with Russia, whom unfriends have converted from a friend to an angry and jealous rival and who is ever so forward in the linguistic field.

But if the model Persian dictionary have its difficulties, far harder will be the task with Arabic, which covers incomparably more ground.  Here we must begin with Spain and Portugal, Sardinia and the Balearics, Southern Italy and Sicily; and thence pass over to Northern Africa and the two “Soudans,” the Eastern extending far South of the Equator and the Western nearly to the Line.  In Asia, besides the vast Arabian Peninsula, numbering one million of square miles, we find a host of linguistic outliers, such as Upper Hindostan, the Concan, Malacca, Java and even remote Yun-nan, where al-Islam is the dominant religion, and where Arabic is the language of Holy Writ.

My initiation into the mysteries of Arabic began at Oxford under my tutor Dr. W. A. Greenhill, who published a “Treatise on Small-pox and Measles,” translated from Rhazes —­Abu Bakr al-Razi (London, 1847), and where the famous Arabist, Don Pascual de Gayangos, kindly taught me to write Arabic leftwards.  During eight years of service in Western India and in Moslem Sind, while studying Persian and a variety of vernaculars it was necessary to keep up and extend a practical acquaintance with the language which supplies all the religious and most of the metaphysical phraseology; and during my last year at Sindian Karachi (1849), I imported a Shaykh from Maskat.  Then work began in downright earnest.  Besides Erpenius’ (D’Erp) “Grammatica Arabica,” Richardson, De Sacy and Forbes, I read at least a dozen Perso-Arabic works (mostly of pamphlet form) on “Serf Wa Nahw”—­Accidence and Syntax—­and learned by heart one-fourth of the Koran.  A succession of journeys and long visits at various times to Egypt, a Pilgrimage to the Moslem Holy Land and an exploration of the Arabic-speaking Somali-shores and Harar-Gay in the Galla country of Southern Abyssinia, added largely to my practice.  At Aden, where I passed the official examination, Captain (now Sir.  R. Lambert) Playfair and the late Rev. G. Percy Badger, to whom my papers were submitted, were pleased to report favourably of my proficiency.  During some years