and Botonga; in other words, for the parts between Delagoa Bay and the River Zambeze. The Portuguese libraries would probably supply much information.

The Mozambique Languages.-Of these the chief is the Makua; for which, besides

1. The Makua and Monjou of Salt, known to Vater, there is,

2. The Makua of the Asiatic Society.-MS. Vocabulaire François et Maqououa, ou Recueil de quelques mots de langue Maquouoan. Commencé en Novembre, 1790. This was in the possession of Sir Alexander Johnstone. I have been allowed to avail myself of its contents.

The Zanzibar Languages. Here the information is almost wholly of recent origin. The Zanzibar language best known is the Sowaiel, Sowauli, Suwahili, or Sohili.

1. Salt's vocabulary-Very short. Known to Vater. Used by Prichard. 2. Nouvelles Annales des Voyages. Paris, 1845. iii. p. 287.-Jülg. 3. A vocabulary of the Soahili language from the Memoirs of the American Academy. Cambridge, 1845.-Jülg.

4. Sowaiel vocabulary.-Bombay Geographical Transactions, 1844.

5. The Suaheli of Krapf.-Collected by Krapf in 1844. MS. in the library of the Church Missionary Society. A lexicon of 10,000 words, with grammatical observations, and a translation of Genesis, the Gospels, and the Acts of the Apostles. Of the translation of Genesis three chapters are published, with an introduction by Mr. W. W. Greenhough, in the Journal of the American Oriental Society, vol. i. No. 111, 1847. Krapf's materials have been examined, either wholly or partially, by the Chevalier Bunsen and Ewald. The former notified, at the Oxford meeting, the labours of Krapf in the Sohili, Wanika, and Pocomo languages, and also the existence of a Galla dialect as far south as the fourth degree. The latter has published in the Zeitschrift der Deutschen Morgenländischen Gesellschaft, a paper, Ueber die Völker und Sprachen südlich von Aethiopen.

6. A Sowaiel vocabulary in Brown's Notes on a Whaling Expedition. 7. Sohili of Leigh.-Soailese (Sowhylese, or Sohili) vocabulary, collected by John Studdy Leigh, Esq., on the east coast of Africa in 1837-39. MS. A year ago this vocabulary was kindly put into my hands by the author. Its general Caffre affinities were less marked in its glossarial, than in its grammatical elements. An analysis of the Pater Noster, compared with a tabulated vocabulary, proved this. It was less like the Makua of the Asiatic Society than was expected. Nevertheless, it had undoubted affinities with that language.

8. The Quilimani Language.—A short vocabulary in the MS. last menAllied to, but not identical with, the Sohili.

tioned.

9. The Wanika of Krapf.-A short specimen in Ewald, Ueber de Völker, &c. 10. The Wakamba of Krapf.-A short specimen in Ewald, Ueber de

Völker, &c. The Wanika and Wakamba are conterminous with the Galla on the north, and the Sohili on the east.

11. The Msegua of Krapf.-A short specimen in Ewald, Ueber de Völker, &c. The Msegua inhabit the coast opposite the island of Zanzibar.

12. The Pocomo of Krapf; mentioned, for the first time, by the Chevalier Bunsen at the Oxford meeting.

13. The Msambara of Krapf. MS. Tabulated, and so shown to be Kaffre.

14. The Ukuafi of Krapf.-A short specimen in Ewald, Ueber de Völker, &c. Conterminous with the languages last mentioned. Ewald's paper, which seems to have been written without a knowledge of either the Caffre grammars of Boyce and Archbell, or of the opinion of Prichard on the Sohili, gives a table of these four languages and of the Suaheli. The Ukuafi is by far the most unlike of any. The others are closely allied to each other and to the Suaheli. The Suaheli of Krapf and the Sohili of Leigh coincide.

15. The Mobilian of Herbert.-Taken by Herbert in 1677 at Mobilla, one of the Comoro Islands, between Madagascar and the Continent. Although very short, and in the words of the author a mish-mash of Arabic and Portuguese, it is shown by the following two words to be African and Caffre rather than Madagascar and Malay :

The language also of the island of St. Johanna is African, not Madagascar.-Prichard.

Between the Caffre language and the other languages a broad line of demarcation has generally been drawn.

The peculiarities, indeed the so-called characteristics, of the Caffre grammar are objects of great prominence in African philology.

of

1. The system of prefixes.

2. The euphonic or alliterational concord.

These consist

A. The System of Prefixes.-In the following words,-umtu-person, ihashe=horse, inkosi = captain or chief, isicaka servant, usana infant, umlambo river, ubuso-face, akutya=ford, abantu people, amazwe= words, inkomo=cattle, imiti=trees, the letters printed in italics are wholly non-radical. Adventitious, however, as they are, they occur before almost all the Caffre substantives, and are, practically speaking, inseparable. Without them, the noun would appear like a Greek crude root, i. e. such a form as λoy- opvio-, a root minus its concomitant inflection. Yet the Caffre prefixes are not inflections. They are elements of composition. So necessary, however, is the incorporation of this adventitious element, that new words like priest and pharisee, &c. take it, and become umpriest, umpharisee, when introduced into the Caffre scriptures.

A measure of the importance of these prefixes may be found in the following sentences from Boyce's Caffre Grammar:

a. "Nouns are distinguished by prefixes."

b. "The prefixes are numbered according to the number of the declension of the noun to which the prefix belongs."

c. "The whole of the grammatical construction of the language depends upon them" (i. e. the prefixes).

Verbs and adverbs are converted into nouns by the addition of prefixes.

Certain prefixes have a singular, others a plural power.

The system of prefixes is nearly the same, even in detail, for both the Caffre and Sechuana.

B. The Euphonicor Alliterational Concord.-Several of the syntactic rela tions between two words in Caffre are expressed in the following remarkable manner. The word governed changes its initial letter into either the initial letter of the word that governs, or one allied to it. Thus if the English syntax was the same as the Caffre, we should say instead of

..man mog.

people's leader ..........people peader.

and so on. In this case, the initial of one word would determine the initial of another.

Such is the rule of the Caffre euphonic concord; with this addition, viz. that the prefix is part of the word; so that it is the prefix rather than the word itself which influences the alliteration.

1. The prefixes u and um determine that the words in a certain syntactic relation to them, shall change their natural initial into w,-umtu welizwe=a man of the country.

2. The prefixes i and ili determine that the word in a certain syntactic relation to them, shall change their natural initial into l,—ihashe lenkosi= horse of the captain.

3. The prefixes in and im determine that words in a certain syntactic relation to them, shall change their natural initials into y,—inkosi yabantu= captain of the people.

So on throughout. The letter which the prefix of the governing word requires to be the initial of the word in construction with it is called its euphonic letter; so that w is the euphonic letter of u and um, l of i and ili, y of in and im, &c.

Now if languages were, like crystals, determined by their external form, and without any respect to descent or affiliation, the phænomena of the system of prefixes, and of the euphonic concord, would be characters of great value in separating any language in which they occurred from any in which they did not. At present it is sufficient to say that they do not occur in the Caffre languages only.

A third point in the Caffre language is suggested by certain forms occarring in Ewald's table. In the Suaheli, the Wanika, and the Wakamba, the plural form=men is simpler than the singular form=man.-Singular mtu, mutu, mundu; plural watu, atu andu. Now, not only is it à priori likely that, in certain instances, the expression for individualizing particular objects may be subsequent to the expression for a collection dealt with as unity, but, in two languages, the Welsh and Arabic, the actual phænomenon of a singular formed from a plural has been recognised; just as, in English, the word wizard derived from witch is a particular instance of a masculine form being derived from a feminine one.

XIII. The Hottentot Language.—The course of the Orange River illus trates the area over which the Hottentot languages are spoken. On the south they are conterminous with the Dutch and English of the Cape, by which their more extrenie dialects have been displaced. On the east and north-east they are bounded by Bechuana and Caffre dialects of the Caffre. The Namacqua, which is a Hottentot language, is spoken on the shore of the

Atlantic on both sides of the Orange River. The Dammara of Walvisch Bay is by some considered a Hottentot, by some a Caffre tongue.

Compared with the scantiness of our information concerning other parts of Africa, the knowledge of the Hottentot class of languages, at the date of the Mithridates, was considerable; what was known being known chiefly through the Dutch missionaries and the early travellers.

The glossaries in the Mithridates are

1. For the Corana dialect.-A short vocabulary of Lichtenstein.
2. For the Saldanha Bay dialect.-A short vocabulary from Hervas.

3. For the Saab or Bushman dialect.-A short vocabulary of Lichtenstein. Although the prominent points of difference between the Bushman and the other Hottentot dialects were known to Adelung, its real Hottentot character was decidedly insisted on.

4. For the Hottentot language in general.-Short vocabularies by Sparman, Thunberg, Ten Rhyne, Witsen, Kolbe, Barrow.

5. A Pater-noster from Witsen.

6. A Catechism by Van der Kemp.

7. A Pater-noster in the Corana, from Campbell's Travels in South Africa. London, 1815.

8. The four Gospels in the Namacqua dialect, by the Rev. Mr. Schmelen of the London Missionary Society. Printed by the Bible Society.-Boyce's Caffre Grammar, Introduction.

9. Of the Hottentot we find a few words in Herbert's Travels, a.d. 1677. They are called Souldanian; and were probably taken at Saldanha Bay.

The Hottentot is one of those African languages which has been separated from the rest by a, real or supposed, broad line of demarcation.

XIV. The Nubian Class of Languages.-The Nile from the cataracts towards its source, is the best guide for the languages of Nubia and Abyssinia. 1. Immediately south of Assuan is the country called El Kenuz. Of the dialect there spoken we have a vocabulary in Burckhardt's Travels in Nubia. The Kensy is the north Nubian dialect. It is also a Barabbra, or Berberine dialect. The Kenuz of Nubia is conterminous with the Arabic of Egypt on the north, and with the Bisharye dialects on the east.

2. The Wady Nuba lies immediately to the south of Kenuz. Of the dialect here spoken we have also a vocabulary in Burckhardt.

3. The Dongolah of Cailliaud.-Voyage à Meroe, &c. Paris, 1826. 4. The Routana of Eusebe de Salle.-Journal Asiatique, 1840, vol. x. The Routana, as the name of a language, is probably the same as the word Ertana, in Lyon.-The Arabic name for what is considered a patois.

5. The Noby of Eusebe de Salle.-Ibid.

6. The Nubian of Costaz.-Ibid.

7. Minutoli, J. H. H.-Reise zum Tempel des Jupiter Ammon und nach Ober-Egypten. Berlin, 1827.