loSS T45 v.l CSfpi'nc Hiiiunnti'c (£mnnn?,jfj.$:<3v. The date shows when this volume was taken. To renew this book copy the call No. and give to the Ubrarian. UUi o . icioi Q HOME USE RULES !!iQ V *^ 8. 1951 Gr ^* ^°'*''^ subject to RecaU All borrowers must regis- ter in the hbrary to borrow books for home use. All books must be re- turned at end of college year for inspection and repairs. Limited books must be re- turned within the four week limit and not renewed. Ltx a..4 Jah iM..l<^.. Students must return all books before leaving town. Officers F.houH arrange for the return of hooks wanted during their absence from M i\ R 4 1070 PM"'"" ^Cj P •; tr i-a»*ir "" . Volumes of periodicals tid of pamphlets are held the library as much aa possible. For special pur- poses they are given out for a limited time. Borrowers should not use their library privileges for the benefit of other persons. Books of special value and gift books, when the giver wishes it. are not allowed to circulate. Readers are asked to r*> port all cases of booki marked or mutilated. Do not deface books by marks and writing. , , ililllllillllllllliiiillll 3 1924 092 510 167 (}}. No data uppca' (o exist for the 12). The boundarir.s of Va: Gala. Kin ANTHROPOLOGICAL REPORT ON SIERRA LEONE. BY NORTHCOTE W. THOMAS, M.A., F.K.A.I. Government Antiikoi'Ologist. PART 1. LAW AND CUSTOM OF THE TIMNE AND OTHER TRIBES. LONDON: HAKKISON AND SONS. lillG {Copuriijht.) 7 ^ua-i // LONDON: HAEBISON AND SONS, PBINTEES IN OEDINAKY^TO'hIS MAJESTY, ST. maktin's lane. Jr TABLE- OF CONTENTS. I. — Introductory ."> 11. — DKMOaRAI'IIY 14 III. —Paramount Ciiikk 20 IV. — Rki.ioion ... 29 V. (JULT OF THK DkAIJ 41 VI. — Witchcraft 46 VII. — Satka, Wanka, ktc. ."i2 VIII. — Ritual I'roiiiiutkjns Gf> IX. — Divination, Orukauk, ( ».mknh ... 80 X. — Marriaok... 91 XI. — KlNSlIIF lo:) XII. — Birth, Twins, Circumcision 108 XIII. — BUBIAI lis XIV. — ToTKMIS.M i:52 XV. — SeCRKT SoCIKTlKS u:! XVI. —Law, ( 'riminal ... 1 .-,:} XVII. — Sl.AVFRY l.-,M XVIII.- — InIIKRITANCK, LaNIJ, |)KliT 102 XIX. — b'ARMINC AND ( 'rOPS 172 XX. — TkCIINODOOY AND SclKNCK 177 Note on I1i>tamc.\i. Fkathrf.-*, hy I>r. o. St.\ff . 181 (Ilossary 18.-, Index 189 LIST OF PLATES. Map — Showing Distribution or Tribes, by Major (now Lieutenant-Colonel) C. E. Palmer, D.S.O., E.A., formerly attached TO THE Sierra Leone Battalion, West African Frontier Force Frontispiece Facing page I. — (a) Susu Weaving ; (b) Suspension Bridge ... 11 II. — (a) Timne House; (b) Limba Stone House ... 12 III. — (a) Konten; (b) Timne Girl ... ... ... 20 IV. — (a) Sanko ; (b) Satimaka 28 V. — (a) Koranko Image; (b) Maskers (Timne) ... 39 VI. — Sacrifice : (a) For Health ; (b) For Good Sleep 41 VII. — Sacrifice : (a) For Bad Dead ; (b) For Farm... 43 VIIL— Susu Boy IX. — (a) Atettot ; (b) Sena X. — Timne Man (Sanda) ... XL— Timne Man (Yoni) ... XII. — Timne Woman (Sanda) XIII. — (a) Bundu "Devil"; (b) Circumcision Mask ... 117 XIV.— Graves : (a) Timne ; (b) Susn 129 XV. — Timne Woman (South) 140 XVI. — Yalunka Man and Woman 146 XVIL— Limba Girl 152 XVIIL— Koranko Man 160 XIX, XX.— Pot Making (Sanda) 177 60 84 90 96 110 L— INTRODUCTORY. TiiK colony and protectorate of Sierra Leone lie between 7° N. and 10° N. in latitude, and 10° 50' W. and 13° .10' W. in longitude, with a total area of about 31,000 square miles, and a native pojiulation of more than a million and ar[uart(;r. The area near the coast is, with the exception of the mouutaincjus region near Freetown, uniformly Hat and, in the rainy season, in many jiarts swam]>y. Higher ;^fround is found eastwards towards tiie Liberiaii border, and hill country is (entered soon after cnjssing the Seli on the road to Kaballa, which lies fully 1,200 feet above tiie sea. witli surrounding hills perhaps 800 I'eet higiier still, H(jme of them occupied by Lindia villages built in jiart of stime. The rivers run in the main from north-east to south-west, the j\Ioa, Sewa, Taia, Sell (Kokelle), Kaliba (Little Scarcies), and Kolente (Clreat Scarcies) being the most im]iortant, though rapids often make miN'igalion inipossibh; ncit far fi(im the mouth. The Seli takes its rise not far from the head-waters of the Niger, here kiujwn as the Jolib.i. Except in the hill country in the north-central area vegiHalioii is exceedingly rich, and there are more than 1,000 species of trees and ]ilants, exclusive of riec and cultivated vi'getables. Among the ainmals may be mentioned the chimpanzee, on the Scarcies and in the area between Freetown and the Liberian bordei's, the hippopotamus and pigmy hippo- potamus, t^lepli.ant, liongo, eoli, bush buck, and a number of duiker; leopard and many kinds nf cat, dwarf bull'alo, and wild pig. Anion'.,' the birds, guinea fnwl, francidin, greater and lesser ]ilantain eaters, bustard, ami many kinds of duck are found ; snares are set for birds, especially near the marshy areas in the south, where spur-winged geese are plentiful. Snakes are abundant, but apparently not dangerous as a rule, though the spitting cobra is not uncommon. Scorpions and land crabs are found, and tish are plentiful in the rivers, though until recently only eleven species were known from the whole area. There were comparatively few opportunities of ascertaining facts as to the prevalence of disease, but the natives do not suffer to the same extent as the Nigerian peoples from indolent ulcers. The various tribes showed a marked difference of character in respect of the readiness with which they submitted to treatment. Except in Susu villages it was very rare for patients to accept an invitation to come for medicine ; but Susus were everywhere ready to come forward, even when the tribe among whom they resided showed no inclination to do so. Old men seemed far less frequent than in Nigeria, but no exact estimate of age was possible in the absence of historical events of known date as a starting-point. There can be no doubt that the wars, which went on till some twenty years ago, swept off masses of the population ; in one town I was told by an old man that seven of his nine wives had dis- appeared in this wa3^ In physical appearance there appears to be a well-marked difference between the Mandingo peoples and the other tribes. The Susus and the M§ndi are more lightly built than the Timne ; the Limba type is different from either, and they are perhaps somewhat darker. Many of the natives in the south appear to be very capable traders ; in one family, four of whose members received part of the farming capital for trading purposes, more than £2,500 was banked in seven years from the surplus profits. In general, however, the native appears to be unintelligent and singularly lacking in initiative ; it is a rare occurrence to give an order to a man and find it carried out promptly and intelligently, or even carried out at all I, The Limba is perhaps rather .superior to tlie other tribes in tliifi respect. Tiie Mgndi is a better carrier tlian the Tinine , and he seems to be generally more resolute, though he is at the same time more light-hearted. On one occasion Tinine carriers, who were called upon to wade a river in Hood, which was no more than chest deep, gave themselves up for lost, when the M^ndi hammock boys were quite unperturbed. The cheerfulness of the M(;ndi, on the other hand, makes him less jirudcnt. Freetown it.self is inhabited by the descendants of liberated Africans, who, fifty years ago, sjioke hundreds of dillerent languages, as Koelle has left on reconl in his Pol (/ylutta Africana; to-day Yoruba (Aku) and possibly Ibo survive. Of the languages of the Protectorate, Fulfulde (Fula) is spoken by scattered sections of the tribe, covering over a hirge part of the area, who are in most cases sedentary, sometimes in their own villages, and mainly occupied with cattle-keeping. The remaining languages arc Soudanese, and fall into two main groups, preHx and uon-preHx tongues. To the former, which may be called the old groiqi, belniig («) Timne, (/<) Limba (with several dialects), and (r) Jiul.im (Mampa Siierbro), Kriui, and Kisi, which are closely related. The noii-preti.x; languages are: {<() Gola, an isolated tongue on the Liberiau border with no known allinities beyoml tliose existing between all Soudanese languages, and {h) the Mandingo group, of which the following are included in Brilisii territory: Susu, Valunka, Koranko, Kono, Yai, Loko, and Mendi, of which the last ditlers in a somewhat marked degree from the normal Mandingo type. Myndi is and has for some time been swallowing up its smaller neighliours, liulom (Sherliro), Krim, and A'ai ; it is by no means inijirobable tiiat some of the features of M\;ndi are due to the fact that the Mandingo element in the tribe is far smaller than in the other tribes nieutioned above ; this is borne out by the fact that M^ndi has several well-marked dialects. On the north-west of Freetown Buleni, of which the Timne name is Mampa, is being swallowed up by Timne, and is only found in isolated groups. It is certain that considerable changes have taken place in comparatively recent times in the distribution of the tribes ; for Port Loko is now a Timne area, but from its name it is clear that it has been in recent times in Loko hands. The Timne occupation of the Bulem shore is also comparatively recent, if the maps published at the end of the eighteenth century can be accepted as a guide. Even in 1854 Koelle says that Timne territory is south of the "Sierra Leone" river. It must, however, be remembered that the treaty of cession of the colony proves that the Freetown area was in Timne hands in 1788. As to the conditions previous to this period we have little or no information save from tradition, which goes to show that at no very remote era the Protectorate was covered with virgin forest, of which the remains are found on the Liberian boundary and between the Timne and aSTorthern Mqndi areas ; south of the forest lay the Bulom, but tlie forest itself seems to have been mainly uninhabited. One curious fact, however, possibly not without significance in tliis connection, may be quoted. The German word for parrot is " papagei," the English word " popinjay " is from the same root, and cognate words are found in Portuguese and other European languages ; both have been traced to Arabic and other roots, but without any great certainty. The Timne, Limba, and Loko word for parrot is pampakei, and it seems clear that the German and Timne words are genetically connected. We know that words for pine-apple (ananas), tomato (tambatis), etc., have been introduced with the objects themselves ; but there is no reason to suppose that parrots, which are comparatively rare e-^'en in the southern Timne area, and quite unknown in the northern portions, so far as my observation goes, were ever introdi;ced in the same 9 way, slill les.s Ihat they were introduced hy Portuguese or other wliite men, as must have been the case if the Timne adopted a European word. If, however, p.irrots were introihiced into Europe hy some- one who visited tiie West Coast in the MidiJle Ages and brought back some of tlie birds, the puzzle is to ex]jlain why they should have adopted a name from tribes of wliicli two — Limba, and Lt^ko — have never, so far as we know, been on the coast, nor on a navigable creek save at Port Loko, ami ihe third pciiietrated to the coast in all probability long after the word reached Europe. There are other tribes nearer tlie Gamhia which ha\'e a cognate name for parrot, and it may be that it is fr(jm them and not the Sierra Leone tribes that the word is Ijorrowed. In any case no certain inference can be drawn from the facts. Tradition states that the Timne came from the east, and not only have thdy isolated the Lgko from the ]\Iendi, sub- sequently flowing round them In the wi/sl, but the Limba mass, north and east of the northern Timne, has the ap])ear- ance of having been imshcd on one side by an incursive ]ieople, but the fact that the ilaga and other tribes speaking Timne dialects aie in Erencli territnry suggests that the jtcople came frfnu the west; pi'rha]is oidy the chief came from the east. As regards the old grouji, Lind)a is delinitely a piclLx language, as the following forms show: — Huujiilur. riiinil. Kiujl ts]^ hutiti tatiti tooth kutai hatai f(jot to hate hiwl huiak inaiak stone wall mbali slave but names of animals, with few exceptions, seem to have a suffix plural : — SilKJular. J'liira/. Eiii/hx/i. kampa kamjiafi elephant kusa kosefi pig 10 The principal prefix is ku, hu (ta, na); fo (fca), fa (mp, 11 a), w (b), are also found. The situation as to Timne is perfectly clear. The prefixes serve to form the plural, and at the same time indicate whether the noun is in the definite or indefinite form, i.e., whether it corresponds to the noun with the definite or with the indefinite or no article. As regards Bulam, Krim, and Kisi, the situation is less clear. There can be no doubt that the languages are closely allied in vocabulary, as the following specimens show : — Biildin. Krim. Kisi. KnglisI rin Qde yinde hair pal depande ■ pal en sun pan yipan pange moon mwen m^n mQiidan water kil boxi keyo house matuhm maieoi kgiyan fat [pie] wis visio meat kuluii bilih kulin goat can kucan kiude tooth mQ kaniQ melin breast su kusu so finger kwen lekin knife taun ho tin eye Bulam forms the plural in some cases by prefix, e.g., can, ncan, tooth; rin, irin, hair; rok, iirok, grandcliildren ; and the same is true of Kfim, which is obviously a prefix language in the above list: — tanye, munye, ear; kuca, niuca, tooth; kusu, musu, finger. Even Kono seems to have the prefix plural in a few forms, e.g., moya, eya, eye. When, however, we turn to Kisi, the plural is formed by suffix or change of vowel in the last syllable: — liotin, hotan, eye; kinde, kindon, tooth; ba, balai'i, hand; keyo, keraii, house ; bQiigo, b^ugulaii, foot. It seems, however, evident that there is no prefix change to form the plural ; yet the forms yinde (hair), hotin (eye) Pj.ate I. susu WKAVixc;. See ])ai>e ll'. susi'ENSiDN KKiUGE t)v CKKErEUs. See ] iai>'e 1 2. 11 wlien we compare them with the forms in the other laii;,'uagen suggest that prefixes are not unknown. In some cases the Kisi suffix change seems to be clearly a change of form in the determinative do, kinde, kin do, tooth ; yomdo, yomde, tree; but in the case of ba, Ijalaii, hand; keyo, keran, house; kamao, kaniani, elephant, we have the addition of !(a), r(a) or n, which apparently indicates the plural. On the whole it seems probable that^ Kisi, which is isolated among the Mandingo grou]), has lost its old prefix methods and adopted sulfix change, in a certain number of cases oidy, as a means of indicating the plural. Articles, or forms of the noun taking the place of articles, are not a normal feature of Mandintio lansuaifes. It is therefore worthy of note that Mtjndi, and Li.^ko, have suffixes i with, in Myndi at least, different f(jrms fur the plural to indicate the delinite ami indefinite forms of the noun. Kisi appears to use do, f), with a jilural iii; Bul.im and Krim have a form de, which is apparently detcnininatixi-, while Limlia and (jola use yo, ho; the Limba plural in ii is perhaps connected with the form ni. T(jnes play a consideraiile jiart in Mcndi, Limlia, and ]'>ul,)m; and their importance in Mandingo languages inay be regarded as probable. In Timne, on tbe other hand, owing, no doubt, largely to the develi>i)ment of prefixes, which vastly dinnnish the possibility of homonyms, they play a very suljurdinale role. The small dialectical dillerences in Timne may point to its being, in its ]iresent form, cjf late origin, which no doubt favoured tiie di'o]jping out of tones. The mode of life and native products of the triljes, of \vliate\'er grouj), sIkjw singularly little \ariation over the whole area. The Yalunka and Koranko are peihaps the most diversified as regards manufactures, tlmugh the Jvorauko in-oduets on sale in Freetown are confined to a small area of the Koranko country. 12 Of important ethnographical features the xylophone (Kor balany§) is confined to the Koranko, the loom (Plate I) to Susu, M§ndi and Limba with few exceptions, pottery mainly to the Mendi, though this is due principally to the introduction of European pots, mainly of iron. The general form of the house is everywhere the same, save where the rectangular house has penetrated, coming from Freetown. It is circular with, as a rule, small rooms outside the main wall, but under the main roof. The thatch is of grass tied down on poles secured to each other by a series of circular rings (Plate II) . The substructure of the walls is of wood, upright poles sunk in the ground with liorizontal pieces to give rigidity. The portion of the house outside the main wall which is not taken up with the kohko, or small room, is often fenced with a low wall and forms a veranda in front and behind, which is sometimes on a level with the ground, sometimes raised above and approached by steps. In the Limba country near Kaballa is found a feature very unusual in West Africa — the use of stone in the con- struction of houses (Plate II). At present there are no data to show whether this was sporadic or derived from some other area ; possibly the use of sim-dried bricks at Falaba may have suggested the idea ; but it is more probably due to the scarcity of suitable material for house-building on the tops of the hills. The double gong is characteristic of the Limba. In this area we also find large mud rice-bins built inside the house, two or three feet in diameter, and sometimes as much as seven feet high. Native suspension bridges (Plate I), in use on the Sqli and other rivers, are perhaps of Koranko origin, though they are also found in the Timne country to the south of this tribe. The hammock and sling are in general use among all tribes ; loads, especially rice, are carried with a pack and head band. Secret societies tiourisli, especially in the M§ndi and adjacent Timne areas, but the Susu and Limba have important I'J.ATK II IKirsK I'.III.DINC AT MAI'dlM. Scc|);i;:c 1 L'. INK IKH'SK (I.IMI'.a; AT VAKALA. Scr [ml; 13 societies. The woman's society, Bundu, is not known to all the Limba, nor is clitoridectomy practised west of Kahalla ; but there is nothing to show by which tribe it was introduced or how it originated in this part of West Africa. According to a MS. of Schlenker, who was in the Tiinne country si.xty years ago, Bundu was learned by the Timne from the Mgndi. Circumcision appears to be universal. Various dialects of Timne are distinguished but the diH'ercnces are small. The nortlicru branch are known as Saiida Timne; it was in this area that must of my enquiries were made. u II.— DEMOGRAPHY. In order to ascertain the proportions of the sexes at birth and in mature life, and to obtain information as to the effect of polygyny on fecundity, the sex ratio of the first- born, the relative mortality of males and females and otlier matters, genealogies were collected giving details of the families of over two hundred and seventy men, including one with fifty wives, who was himself the son of a man who had sixty wives and one hundred children. Fifty-three daughters in these families had gone to husbands, but there was no information as to whether they were monogamous marriages or not. There was, however, a tendency to omit or overlook their dead children, and the same was true in a more marked degree of the information about older generations of my informants' families. These data are therefore not as a rule included. In addition to these genealogies a few villages were completely counted and random samplings were made of cliance assemblages of men at various places. For my genealogies were mainly derived from sub-chiefs and were apt to show an undue proportion of polygynous marriages, and might introduce other errors into the data. On the whole, however, it was found that the systematic census could not be carried out with success even with the support of the paramount chief ; in more than one place the information vouchsafed in his presence was plainly erroneous and deliberately falsified. In other cases the paramount chief refused to give any assistance in the enquiry. On the wliole, therefore, the data collected in this way were in bulk con- siderably less than those obtained by the genealogical method, and the reliability was inferior ; the glaringly erroneous data have, however, been omitted. On the whole, the two sets of statistics show- such close 15 agreement as rei^ards the sex ratio in the total birtlis that tlierc can be little donljt of their relialjility. As might perhapH be expected, the mortality, both among males and females, was in'gher in the general count tlian in the genealogies. Generally speaking, in the genealogies there were 422 male Inrths to 2o8 iemale, a ratio of 100 to 61 ; in the census 294 to 206, a ratio of 100 to 69. Males surviving were to females surviving in the latios of 100 to .5.5'6 and 100 to 04. Tal, 25, 24, 3'9 (mainly owing to two families, one of 27, one of 17), 3, and 2-0. It has, however, been noted above that the census figures were demonstrably unreliable in some cases, and it is by no means improliable that while some unduly diminished the numbers of their cliildren, others unduly magnified them. It seems clear from the census that males are slightly in excess of females. The results show 860 living males of all ages and 348 females, exclusive of casual lodgers in a house with only remote rehitionship to the head, or, at any rate, no direct relationship of descent. These include a certain number of women who have lost their husbands, and if they are included they would be nearly, if not quite, counterbalanced by the males, more especially younger brothers, who had lost their fathers and were without homes of tlieir own. The age of marriage being much lower for the woman, the number of females living under corresponding circumstances is also much lower. On the whole the proportion of twin birtlis appears to Ije c 18 exceedingly low. Twins are not regarded as unlucky, and there is no reason to suppose that any twin births were concealed. The fact that in many cases special names are given to twins, such as Sento and Sino, and that the child next after twins is often called 'Bfese (to be distinguished from another name similarly spelt but with different tone — 'Bte(^), makes it comparatively easy to detect the presence of twins, even if the actual information is erroneous or defective. Excluding the descendants of Seni Kabia of Magbile, only six or seven cases of twin births are recorded in the genealogy, a far lower proportion than appeared to prevail in Nigeria, where twins are, or were, systematically exposed by many tribes, and, prima facie, the twin-bearing stock correspondingly depleted. In the family of Seni Kabia, on the other hand, twins were far more numerous, though not perhaps exceptionally so in comparison with the normal European stock. Seni Kabia was himself a twin ; among his children were two pairs of twins ; and one of his children, not himself a twin, had also two pairs, as to whom no further information was obtained. Fode Kabia, son of Seni Kabia, had by his first wife three sets of twins among his twelve children, all but one of the twins being males. One of these twins was himself the father of twins, born prematurely. One other son of Fode was also the father of twins. There were in all ten pairs among 250 names recorded in the genealogies of the descendants of Seni Kabia ; only six or seven were recorded, on the other hand, among the 750 names in the remainder of the genealogies. In Tables I and III square brackets show the number of dead wives, round bracliets tlie number of unfertile wives. 19 en w o _• •S3[-BUt9j[ -r CO CI -f 00 32 Cl H -KOI-EJ^ OS CO CI CO CO CI CI Cl 1—1 •8JIM. q^9 _ o 1 "^ CI •-' i ^ ■ajiAV qif. •* CO o 1 '^"~ ■ajiAv pjg •-' ril CO O 1 00 ■a}]jA pus ■* lO CO ■* 1—1 1 *■" ■ajiAi isi •pB9p 88[BUI9J£ ead. Alive. Dead. 1. Three of K'w father's wives. Also Baki (hoh). S^i 1 2 1 1 2.(1) z (i!) 1 - a .'» (3) 2 2 1 1 Also brother's wife, ehildles.s, in No 10. 3.(1) H -2 1 (i) 1 — 3 1 4.(1) 1 2 1 1 •) (^) (3) (iriarried in preceding year). V M ji )» Also Basi (si>ii). (1) — — 1 1 — (i) 1 — — — 5. (1) -— — — (2) (niarriecl piecetling year). A1m(] one Ijrotlier. (1) 1 — j — 1 0. (1) (a wicliiw) Also his mother. — 7.(1) — — 1 — . — (2) — 3 — 3 3) 2 11 — (4) (recently married and pregnai It) 8.(1) — — — — (^) 2 — 1 .5 Also Morlai (son), one wife, m , r lildien. 9.(1) — — I — -_ (-2) {d ead) -- 2 — i^) i> — — — — 7\lso mother and small brothe r. 11.(1) — 1 — — 12.(1) 3 1 — — (2) — — — 1 (:'') 1 ' 1 1 — Also a. w idipw, Binki, in this hous e. ~ I'utitlH l.'j IIUMI, 27 2-1 1 13 21 ]f) (+ 2 dead) and two mothers, two bi .tl crs, one br. tiler's wife, and four w id. >ws. 24 From the close relationship of all the men of the village, it might be imagined that it was founded by Pa Woso's grandfather, but this did not appear to be the case, for it was said to be the oldest in the district. Possibly the explanation lies in part in the fact that four of the men had died in the previous months, and four more were gone away to work. This accounts also to some extent for the large proportion of wives — nearly two per adult male exclusive of the widows. III. -PARAMOUNT CHIEF. The kande or paramount chief seems to be a comparatively recent tignre in many areas; some, it is true, trace their power back to Bai Farama (or Farania Tami), who lived perhaps four hundred years ago ; but in many cases the chieftainship seems to go back only a few generations, and the original chief is said to have gained his position by his wealth ; for in olden days it appears to have been the duty of a paramount chief to compose ilitferences by liberal presents to both sides. Where two tribes were fighting, a chief might spend £15 to £20 and send money to both parties. If they agreed to stop hostilities, an oath was taken and a cow, given by the chief, was sacrificed This was divided into tliree parts, one for each of tlie contending parties and one for the chief. In other cases, especially where there are now two "families " (i.)\, abuna, clans) which share the succession to the chieftainship, there is a tradition that the second family gained its position owing to assistance given in war to the original family. The rule of succession is, of course, not in the direct male line where there are two or more houses, which sometimes represent, not original clans, but des- cendants of the same male ancestor; the term for'iiouse" in tliis case is kunte. In some cases at least the two-clan rule is traced back to the fact that a sister of the original chief or one "f liis successors was the mother of a man elected to the chieftainship. Under the rule of e.xogamy this necessarily involves a change of clan. It is, however, clear that the simple dying out of the oi-iginal family or the youth of its eldest male membeis when the time came for them to succeed, is not an adequate 26 explanation of succession through the female line ; for in some cases there is a record of a chief's daughter helping him and being the overseer of his house. This suggests that she probably attained some authority and was able to influence the election. In some cases the two houses are actually of different tribes ; at Kamalu there is a Loko and a Timne line, though it must be remembered that, owing to the suzerainty claimed by Brama Sanda, Kamalu was never recognised as a full chieftaincy. The original chiefs were by no means necessarily of the same blood as the people whom they governed. It is, indeed, not quite clear how far the country was populated, nor how far the travellers from the east brought their own people with them ; but tradition makes it clear that Koranko, Mandingo, LqIco, and Liniba houses are now among the Timne chiefs. The paramount chief is supreme in his own district, and can in theory decide law cases as he chooses ; if, however, his decisions are glaringly opposed to recognised law, the aggrieved suitor has, at least in theory, the power to go to another chief, of repute as a judge, and, after paying a fee, state his case to him ; the chief would then send to the chief who tried the case originally, and request hun to remit it to him for re-trial ; this, of course, in order to secure the presence of the other party and the witnesses. It appears to be held that a chief cannot, without loss of reputation, refuse to allow a fellow-chief to review his decisions ; but whetlrer a suitor would gain anything in the long run by appealing against his chief's law is quite another matter. His authority over his sub-chiefs appears to be almost absolute. The customs with regard to the election, crowning, burial, etc., of the chief ilifler widely from place to place, and no generalised account of the matter can be given. Many of these customs are regarded as secret, and it was impossible to check the information, which appeared to be given in good 27 faith and with the tacit consent of the chief ; in more than one case, in fact, the chief's son himself gave me the facts, eviilently on his father's behalf. The chiefs are subject to many ritual prohibitions over ami above those incumljent on the mass of the people (see p. G9) ; ami it is a matter of some interest to discover how these arose. A natural idea would be that those chiefs who came from other tribes brought with them their tribal customs and retained them imcbanged ; but of this there does nut seem to be any evidence ; on the contrary, Koranko chiefs in their own country are singularly unhampered by ritual restrictions. It appears far moie probable that the prohibi- tions were developed, at least in part, by the contact between a foreign chief and an indigenous people, precisely as tin; contact between two races has a tendency to cause the ilevelopment of secret societies. In tlie Sanda country, corresponding to the subordinate position of the cliiefs, the niasani and the customs generally are of minor importance. At Kamalu the choice of a new (•hief, always from the utlur house, ix., from the alternati\e line to that of the last chief, appears to rest with the men nf sixty or over. A man of tiie second liouse appears to act as adviser to a chief of the first bouse. 1'iie chief is known ;is ]5ai Sanmra. When the time comes for crowning, the chosen man is shut up in the kanta for six days or more; his wife cooks for him, and the important men may also see him. The house is "opened" for the chief to come out liy the orok, or regent, who is often liis sister's son. The cluef is taken to tlie grass field and they name liis clan, and ask the other clansmen if they are glad ; thereupon they put con- tributions in a calabash, and a moriman puts a wliite cloth on the chief's liead and takes the money gi\en. Tlie enil of this cloth must liang down to the cliiefs waist from the liack (if his head ; for a sub-chief it is ou the riglit side. When he sends for a town to work for him, tbej' work for one day ; four d;iys later another town comes, and so on. 28 In the remainder of the Timne area we find the para- mount chiefs proper, who are divided into Poro and Kagbenle chiefs (see p. 143). All, or nearly all, are subject to special ritual prohibitions, the origin of which is obscure. Although tradition says that the chiefs came from the east, and some, at least, were Korankos, ritual prohibitions of this kind were not found among the Koranko chiefs visited. In some chiefdoms there is a man who represents the " chiefship krifi." At Mamaka he is called Sanko (Plate IV) ; Sanko and the chief, Satimaka, must be in separate houses ; like the chief, he may not go where bundu implements are kept, nor where there is a new-born child. It is significant that at the chief's death his Sanko retires and is replaced by another man after offering a sacrifice. Sanko wears a helmet of leather surmounted by a tuft ; the face is of brass and there is a brass plate behind ; strips of leopard skin are attached to the base, and over the skin is fibre that reaches to the waist. He has fibre ruffles round his wrists and net anklets with fibre tops. Four sticks tied together (bQnk^loma) are in his hand ; they are the chief's staff ; in point of fact the staff actually used by the chief is quite different, long and forked at the top. The chiefship mask of Magbile is known as arQn arabai ; like Sanko the wearer cannot come out when the chief is dead ; the mask is kept in the chief's house. The dress is formed of skins, and he has palm-fibre trousers. When he goes out to walk through the land he carries a broom and whips to fiog people who do not come out when he dances. He can judge cases and pay the money received to the chief. I'l.ATK \\ 29 IV.— KELIGION. Altliougli the coining of Mohammedunisni doubtless modi- fied })rofoundly in some directions the traditional belief and customs of the natives, especially in the direction of decad- ence, without any corresponding influx of new ideas or rites to take the place of those that fell into desuetude, it does not seem difficult to descry the main features of the pre- Mohammedan religion. Tljis did not diH'er very widely from what is found, with larger or smaller variations, in many (jther jiarts of the negro area. (1) The main deity, known as Kuru, or Kuruniasaba, appears to have been a sky god ; he occasionally receives sacrifices, and this is possibly a Mohammedan innovation, for in the prayers that are the main feature of the satka (see p. 52) Kurumasaba is clearly regarded as the equivalent of Alia. Though Alia is doubtless the deity invoked and approached in the mosque, where the niQri man's influence is supreme, it is Kurumasaba who is addressed, even by the good Mussulman, as soon as the literary influence is lelt behind. The mere fact that the satka comprehends so many and so diverse elements is a proof, if any were needed, that the I'cligious life has been little influenced by Mohammedanism ; liut the name satka is clearly due to its influence, and may have been extended to cover many rites not originally embraced under one name or regarded as identical in their nature ; we need not, however, suppose that Kurumasaba was originally called upon in all these diverse ceremonies. In addition to Kuruniasaba, we find at least one, and, ]ierhaps, more than one, shadowy figure that suggests a heathen pantheon in former days. It was formerly the 30 ' custom, and the practice still survives sporadically, to weep ' for Kumba at the beginning of the farming season. A long shed was made near tTie town and hoed with hooked sticks instead of iron hoes ; then rice was planted in the ground thus hoed. All the little children in the town went out repeating, " We cry for Kumba ; they are planting his rice to-day." Some of the people in the town danced and sang : " Abok Kumban o, abok Kumban o" — "We cry for Kumba, 0, we cry for Kumba." No one did any other work on this day. The rice was left uncut, for it was Kumba's ; he was said to be a very bad man who spoiled all the rice in the world, and they had to plant his rice, they believed, when they began farming. This account of a custom now almost forgotten suggests that Kumba was a vegetation god of the type of Adonis ; nothing could, in that case, be more accurate than the state- ment that he owns all the rice in the world ; and when the custom began to fall into desuetude, the belief might well take the form that Kumba had to be propitiated in order that he might not spoil the rice, his rites being regarded as on all fours with those that are directed towards preserving the rice from the krifi and animals. In a neighbouring town, though of another tribe, Kumba was mentioned in another connection ; when no rain fell, all people who had farms went out singing, "We Kumba." The meaning of this was unknown, nor was anyone able to inform me who Kumba was, though it was suggested that he was a witch. The rice gardens in this town, which belongs to the Lgko tribe, were made in the same way as in the Sanda (Timne) country, and the additional information was vouchsafed that the hooked sticks used as hoes were hung in tlie roof of the long hut. The custom appears to be known asTubaiiga; it was associated in their minds with their " old people who died," for they interfered with the farming ; the hoes they used 31 were t?ie hooks hung in the roof, and it was Ijelieved that they would use them again in cultivating the rice sown for them under the long roof. It was added that rice had to be offered on the graves, for otlierwise dead men would catcli the hoes, and there wo\ild be no good rice. The rice sown in the huts was left uncut ; it was masom. These account.s, perhaps, hardly add to the probability that Kumba was a vegetation deity, though it is true that powers (jver rain are iiscribed to these beings in European folk-lore, and it is a Lyko belief, apjiarently, that Kumba withholds the rain, or that an appeal to him will cause rain to fall. It might lie argued that the rice gardens were connected with the cult of ancestors when the Kumba belief fell into the background ; but it might equally well be argued that the belief in Kumba's connection with rice was secondary. This explanation seems more probable in view of the fact that farther south in the Timne country Kumbah was not mentioned in coniiection with the custom, which was called Atobaiikere. Two small huts were built with a path between, and rice was sown in this and left to fall. It was universal .some fifty years ago, but is now obsolete. (H) Below the main deity nr deities ccime, as might be exiiected, a mass nl' minor sjiirits. (a) Some, and these form by far the majority, are name- loss and known only by the generic term krifi, wliich in some triln's are H'.). ^rAsKl•:l:s at m.\k\: aikin atoma: nkmaxkkka. Sre pa<'c 148. 39 stones ; they receive a sacrifice of rice bread and kola ; every five years in October a sheep is offered. The good yina give children, rice, etc. A yin comes with a strong wind and can knock a man down. Only certain people can see them ; a man who sees a bad yin goes mad or falls sick : they walk in the big bush at midday. They can paralyse a man or make him dumb ; a woman who dreams of bad yina becomes sterile ; a yin sets fire to a house sometimes. To keep away the bad yina, the Koran is washed and the water poured on sand, which is scattered round the town. A child born with teeth is a yin ; it is carried to the river and put in; if it is a yin, it goes down the river in about half an hour; if not, it is a human child, which is buried iu the river if it dies. A child with a long head is also a yin. Koranko. — Carved stone images (Plate V) or heads are not uncommon in this area ; and they seem to be regarded with veneration. At Yarawaya is a carved female head with closed eyes, standing perhaps some six inches high. It is said that twins have s])irits behind them and there- fore they may not be with people who are reaping or tliresh- ing rice. One bunch should be cut and put on the road leading to the town ; the twin takes this and says : " I have taken ours ; those who are behind me don't take from the farm " ; then the crop is safe. Witches appear to be less feared in the farm than malicious spirits. Three balls of rice bread are made after reaping ; one is put where the rice will be heaped, one on the road to the town, and one on a lieap of sand ; these are for the krifi. Straw and pepper are burnt on the road on the day on which they thresh the rice. Yalunka. — The word used for krifi is n'in^na. Loko. — Tlie word for krifi is ngofo. Limba. — There are various names for God — Kanu (Safrgko Liniba), Masala (Sella), Masaranka (Tohko) ; but no informa- tion was obtained to show the precise position of this deity. Ninety-five per cent, of Limbas are said to be pagan, and it 40 is significant that the different names mentioned above are in use. Various wali are known. Sokoso shoots people behind the shoulder, like Aronsoh, and to cure the resulting craw- craw a leech must take the shot out. TintryomQ is very long, with a head like a duck and bells at the end of its tail ; its scales are mica. Either this wali catches a man, or the man catches him, according to which has the better eyes ; and its captor becomes rich. Stones are kept in boxes and people cook for them ; they say they take them that they may get good crops. When- ever they see a nice-looking stone, they take it. None of these stones were shown to me ; but if the account is correct, we have here a practice not very remote from that which is commonly called fetishism. One way of procuring rain is to throw water on a wali. Only witches can appear as ghosts ; if others appear it is " only a dream " ; they come to ask for sacrifice. They live in kat'iQ, which must be in the ground, for the bodies are put there. Plate \L -^ACIMFICK (MAT) Fdl; HEALTH. %^' ' 8A('i;iiTii-: KDi; coon sr.EEi'. 41 v.— CULT OF THE DEAD In dealing with the Hubject of the krifi (p. 3.'>) it was mentioned that there was no clear line of demarcation between the non-human and the human spirit. It is quite possible that there has been a certain amount of transference from one category to another; thus, we have seen that the tambara antqf (p. 34) live in big trees ; at Makuta a heap of stones under a cotton tree was known as masar ma ambaki (the stones of the old people), and sacrifices were offered there. This is allied to the cult of ancestors by the name ambaki, and we may suspect genetic relations. Pcjssibly as the site of the town has been changed, a new boromasar was founded, so that two existed, one