for the other Governments of India would be of immense value, not only to the postal service, but to every department of government, and to both the scientific and general public. Broken up into shorter alphabetical lists for the different districts or zilajat, they might be employed in tabulating the census returns; they would form the best of all bases for an Index to the maps of the Great Trigonometrical Survey, and for a guide in correcting the spellings on them; and for office reference at all times. The English spellings would require no diacritical marks except on the long sounds of a, i, and u―provided the name were also given in vernacular characters, which is easily secured in the first compilation, is an effective check on misprints in the Roman character, and a sine qua non for perfect accuracy as well as for etymology. The ease or difficulty in securing correct spellings in such lists centres in the ability and energy of those employed in the different districts to prepare them for the press. It is not the village postmasters and pupils of Government English Schools, however, that should be depended on for correct spellings, but the old headmen and Brahmans, who are natives of the districts and know which are the true forms. There is a rapidly developing demand that place-names everywhere should be correctly represented. In Ireland no expense was spared by Government in employing competent scholars to aid the Ordnance Survey in securing the correct spelling of Irish place-names; and for the revision of the maps of the Highlands of Scotland the Director-General of the Ordnance Survey has asked the co-operation of the Royal Scottish Geographical Society with respect to Gaelic and Norse place-names. But a far wider movement, to the end of securing accuracy in the place-names of all countries, is elsewhere taking shape, and must rapidly spread. Nor have the Indian Government Surveys been of late altogether indifferent to the matter. The later maps are greatly better than the older ones in this respect, showing that care has really been given to the spellings. But the results of long-standing neglect have yet to be looked to. To re-engrave a hundred or more of the large sheets is not asked; that may, and perhaps for other reasons must, be done by-and-by. Meanwhile there are no indexes of geographical positions. These are greatly needed, and should be prepared in the survey offices; while, concurrently, lists for each pargana, taluka, and district should be collected, giving the correct vernacular and Romanised forms, from which the mapspellings would ultimately be corrected. These lists, combined with the positions, would serve the important purposes already referred to; and, when a map had otherwise to be re-drawn, they would be ready to hand. Such an index for the whole of the survey maps, and slightly expanded to form a condensed but complete Gazetteer of India, giving the geographical position, population of towns and villages, heights of mountains, and ever so little other information, would be invaluable for office and general reference both in Europe and India. It surprises one that such a work has never been prepared. Thacker's Reduced Survey Map of India (1891), prepared by J. G. Bartholomew, F.R.G.S., already referred to, is on the same scale as Keith Johnston's Royal Atlas map-that is, of an inch to a degree of latitude. It is beautifully engraved, but without the hill-shading, and has been constructed with special reference to Hunter's Imperial Gazetteer,—every place mentioned in that work being shown on this map. It thus contains ten thousand names,1 all spelt as in the Gazetteer, and forms a necessary and valuable accompaniment to it. The index, which is very full, presents the same spellings, and thus helps to disseminate and perpetuate their errors. In the interests of science it is earnestly to be hoped that when a new edition is required, these mispellings will be revised, and the correct forms given at least in the index. gazetteer or an index, it would be well if the description or position were always given under the correct form, and the inaccurate only with a reference to it, and perhaps only when in the alphabetical arrangement it stood more than a few lines from the other.2 In a The Imperial Gazetteer itself, however, is only provisional, and it differs in the place-names from the provincial works. When it comes to be revised (as it soon must), if the work is intrusted to a scholar, the new-fangled Hunterian place-names must be displaced. No better method, perhaps, could be devised than that employed by Professor H. H. Wilson, in his Glossary of Indian Terms, in which he arranged every article alphabetically under its correct transliterated spelling, and at the end of the work gave a list of the variants with references to the proper pages. This enables the reader to find at once what he seeks for and shows him-if it does not also encourage him to use the correct forms. Accuracy would not be at all a serious or confusing element, as some thoughtlessly suppose. Many names, long and widely known, have been, and still are, spelt in various ways, some nearer the correct form than others, and the adoption at once of that form could hardly in any case lead to a mistake. Katiwar and Kattywar have lately all but disappeared in favour of Kathiawar, and many names have in recent years undergone even larger changes for the better, while there are hundreds of names that would suffer less change if spelt correctly. Further, why should we accept the newfangled "Narbadá" rather than the proper Narmadá, or object seriously even to the restoration of Mathura for Muttra, Kachh, Bet for Beyt, Bharoch 1 Notwithstanding that there are about 1700 more names on this map than on Dr. Keith Johnston's, it would appear, from a comparison of the first three columns of the indexes, that about a sixth of those on the latter are not found on the new map, nor, consequently, in the Gazetteer. A proportion of these may be due to differences of spelling, and some few may be places of very little note; but others better deserve a place than perhaps a tenth of the names entered in the Gazetteer. There are about 5200 names in the index to The Student's Geography of India, by Dr. George Smith (1882). 2 For library and office use there was published, fully thirty years ago, by E. Stanford, The Atlas of India; revised by J. Walker, Esq., selected from the maps published under the superintendence of the Society for the Diffusion of Useful Knowledge (N.D.). This contained India in twelve sections, to a scale of 35 miles to an inch, with an index map, and thirteen others of Afghanistan, Persia, Egypt, China, Japan, etc. This Atlas was a second edition (the first containing India only), in which the maps had been in whole or part re-engraved. Owing to the scale, fulness, and clearness of the maps, and convenient size of the volume (9 inches by 141), the work was very useful and popular; even yet, though out of date, it has not been superseded; and a new Atlas of similar character, the maps being to a scale of about 30, 32, or 35 miles to an inch, equally clear, with better spellings and an Index, would be most valuable for general use, and to officers in India, as well as to all tourists and students. for Broach or Baroch, Bor-Ghát, Katak for Cuttack, Jaypur, Qándahár, and the like? If only employed in maps, in gazetteers, and handbooks they would be readily accepted; and, in the case of less notable places— post towns, railway stations, rivers, hills, etc.—if they were spelt correctly on post-office stamps, time tables, district reports, census and revenue returns, as well as on maps, they would be at once accepted and gain currency. The Great Indian Peninsular Railway, not very long since, changed the spellings of nearly all the names on station sign-boards to the transliterated Marathi spellings, and no inconvenience was caused. When once a correct form has been current for a few years all aliases can be dispensed with, and will naturally disappear. Government could easily bring all this about through the means and agencies already indicated. THE ORTHOGRAPHY OF AFRICAN NAMES AND THE PRINCIPLES OF NOMENCLATURE. BY W. A. ELMSLIE, M.B., C.M., Livingstonia Mission. NEARLY fifty years ago the question of the introduction of a uniform orthography for all the cognate languages and dialects of Africa was taken up and discussed in connection with the work of various British, American, and Continental Missionary Societies. The Standard Alphabet of Lepsius, the use of which would have harmonised all publications issued under the auspices of the various societies, was recommended for adoptíon, but was not generally accepted. This was at a time when African philology was in its infancy, and when the most favourable opportunity presented itself for introducing uniformity. It was allowed to pass unimproved, and now, instead of a uniform system of orthography within the field of Bantu languages, we have four or five different systems in use. Since African philology has assumed a leading position in connection with the development of the country, these variations in orthography present themselves as confusing and retarding elements in the work. These elements will be more apparent year by year. Hitherto, for the most part, only one missionary society has been working with a given language, but now that Africa has been divided among the various powers, and the most natural boundaries have not in every case been followed in this partition, different nations with different orthographical systems may soon be found working with the same language. As an example, we require only to think of the Nkonde region at the northern extremity of Lake Nyasa, which has been so unnaturally divided between Britain and Germany, where henceforth their missionary societies will be labouring with different systems of orthography in the same or allied dialects. But the want of a uniform orthographical system is not regrettable merely because of its relation to missionary effort. The various societies are not so likely to feel it as the African himself, and as the various commercial and international interests will. We cannot but see that, when the friction of the present operations of the Powers has disappeared, there will be a more general community of interest than is now apparent. The African, though he has been educated in one mission, will be found making his home near another, and he will be handicapped by the differences in the languages which have been created by Europeans. The best authorities are agreed that Africa will only be developed by utilising native forces, and, though it may appear a small matter, the want of a standard uniform system of orthography will be found to be a hindrance to international progress in Africa. These remarks apply to the general field of languages, but in the limits of the present paper we must leave the general question and confine ourselves to it as related to the British sphere, and particularly to the nomenclature in connection with geography. Since the time when Livingstone discovered Lake Nyasa, and helped to fill up the blank in our maps of that time, some ten or twelve different workers have been engaged in Nyasaland, and the fruits of their labours have been embodied in published maps of that region. It is not to be wondered at that there should be some confusion in connection with the nomenclature adopted by the various workers. The wonder is that there is not more. But, if at present there is no hope of attaining to a uniform orthography internationally, we may at least endeavour to ensure uniformity within our British spheres, and with this aim the following remarks are put forward, as the present is certainly the most favourable time for united action. A cursory glance at the various maps shows that, while many of the names which appeared on the older maps have been retained, other names have been inserted along with these, on the newer maps, while there is nothing to indicate what relation they bear to each other. There is, in several of the maps, no proper distinction made between the name of a village, a chief's name, or the name of a district. We are, of course, approaching a more perfect knowledge of territories, and, instead of having maps merely showing the general features of a country, we are acquiring more detailed information, and are able to issue maps on a larger scale, which correct the errors of, and supply the features wanting in, the older ones. The orthography of African names does not seem to require revising so much as the system of nomenclature, and yet upon the orthography the correct pronunciation of the names depends, and to the general reader this presents some difficulty, which we think may be easily cleared away. There is no need to adopt a new alphabet, as by means of the Roman characters every sound may be expressed. As Nyasaland is the field more especially under review here, we may say that all its languages are closely allied to each other, excepting the Ngoni [Zulu], regarding which we need say nothing here at present. Having fixed upon the Roman characters, we must next see how they are to be used. Some travellers have adopted certain combinations of letters, and introduced diacritical marks in connection with names, while other travellers have made use of neither in indicating the same sound. Is a uniform method possible and practicable? After seven years' work on Bantu languages we think it is, for all Nyasaland languages. The plan here advocated is not new; it is simply the basis, according to which all names have, to some extent, but not exclusively, been written, reduced to uniformity. We may consider : 1. CONSONANTAL SOUNDS.-All the consonants are not required to indicate sounds met with, as we shall see, but the rule is that they are sounded as in English, except the following: (1) C-The soft sound of c is correctly represented by s, and its hard sound (as in English words) by k. It is therefore redundant. (It is used in Zulu, etc., to represent the dental click). The combination ch, as in monarch, is represented by k, and as in church by tsh, in conformity to the phonetical principle advocated. (2) G-Its sound is always hard, as in give. (3) Q-This letter is redundant, as the English combination qu is correctly represented by kw. (4) X-Is redundant also, as ks represents the English sound of this letter. We are thus able to discard the use of c, q, and x, and are still left with means to express all simple sounds met with in African names. etc. mp, Various compound consonantal sounds are met with, which, however, do not present any real difficulty. They are chiefly a combination of a nasal with another consonant, such as ng, ny, nd, nj, nk, and mb, mf, These have the same sound as in common English words. The sound of ng, as in singing, is met with. Where n is combined with g or k it has frequently this sound, but we meet with the sound where no g or k has to be sounded in combination. This sound has been written ng', but it is of rare occurrence in connection with geographical names, and may be put aside. 2. VOWEL SOUNDS.-The vowels have the open Italian sound a, as a in father. e, as a in save. i, as ee in seen. o, as o in bone. The diphthong sounds will be referred to presently. 3. ACCENTUATION.-Since each letter has only one sound, and as the accent always falls upon the penultimate syllable, the use of accent marks and double vowels indicating where the accent falls is entirely unnecessary. All syllables end in vowels-i.e. are open. If we remember this characteristic of all Bantu languages, we can dispense with the use of double letters, and may always at once place the accent on the proper syllable. This will be evident from the following examples: |