One such caravan, whose leader came from Kankan, contained 534 men and was assembled at Berri Brimayah in 1878. Proceeding westward it came to the coast through the upper Scarcies area of Tambakha to the town of Tagbe on the Melacourie River. Most of the barter was completed there, the inland traders fearing the coast-wise journey to Freetown. Also, one of their commodities - slaves - could be marketed only to the French dealers who were there protected by distance and difficult tidal waters from the British squadrons. Some fifty members of the caravan, however, seeking the higher prices for legitimate merchandise which Freetown offered, embarked in canoes for the final stage of the journey to that town. Here they bartered produce including ivory, gold (worth together $1,040), twenty-two oxen and 2,350 hides.¹

This caravan, though the route which it took may not have been fully typical (cf. Fig. 1) gives evidence of the pull which Freetown's functional growth and development had come to exert on trade arriving at the coast. Already by 1850 Freetown, with its annual exports of £100,000 worth of "African Produce" ² was assuming a dominant position over the many other potential centres along this south-west facing section of the Guinea coast.

It is not easy to assess quantitatively either the extent of Freetown's supremacy or the degree to which the produce shipped to Europe and America was of coastal or interior origin, but some suggestions on the latter point may be based on a sampling of the

¹ Sierra Leone Archives. Ms Route Book 1878-1888, p. 19. S.L.A. Government Interpreter's Letter Book 1878-1880, p. 97.

² The values for "African Produce" are built up from separate commodity totals in the Sierra Leone Blue Books. In some of the later years it is not always easy to distinguish between this category and re-export times.

export data in the Sierra Leone Blue Books. Cox-George¹ has commented on the inadequacies of these data, but it seems likely that orders of magnitude of produce passing through the port of Freetown are correct. He has suggested that the Colony's economy became dominantly "export" in character in the mid-1840s.

Earlier, the principal exports had been of timber (69.2 percent by value in 1829; 58.2 per cent in 1840) but the inland margin of profitability for exploitation of forests was soon passed, the proportion fell rapidly to 10 per cent in 1850, 6 per cent in 1860 and less than 1 per cent in succeeding periods. However, the importance of other local forest products increased. For example, camwood exports increased from 5 to 13 per cent of the total in the period 1829 to 1840 but fell back thereafter. Palm oil, only 6 per cent of produce exports in 1829 approached 20 per cent for much of the period 1840 to 1870 (22.4 per cent in 1850), but declined later as the export of palm products in the form of kernels assumed dominance (Table 1). No doubt the sudden appearance of kernels in

1860 reflects the cession of the Sherbro in that year; and it appears that this area continued to dominate in this trade at least as late as 1899, when 67 per cent of the Sierra Leone total by weight was shipped through York Island and Bonthe. Until human porterage was replaced by other means of transport, kernels were only exploited within a distance of a few miles of navigable water. ²

This was not the case with the lighter, more valuable commodities such as gum, rubber, hides and, of course, ivory and gold. Gold exports are recorded in 1833 (£818 value) and had been made earlier. In 1860 this commodity represented almost 20 per cent of Colony exports, but thereafter one cannot certainly distinguish between

¹ Cox-George, N. A. Finance and Development in West Africa, (Dobson, 1961), p. 143.
² All national values quoted in this and later paragraph are computed from Blue Book data. The 1899 value for Sherbro is from Alldridge, T. J. Sherbro and its Hinterland, (Macmillan, 1901), p. 67.

exports of the raw metal and movements of specie and bullion. It seems certain¹ that all the primary

export gold passing through Freetown came down from the north-east in caravans such as that described above. Hides (worth £10,832 in 1850, £32,881 in 1860 and £46,452 in 1867) and rubber (worth £62,651 in 1880) were also in part brought down from areas outside the present national limits of Sierra Leone, whence came also much of the exported ivory (£4,101 or 7.2 per cent in 1829, £15,589 or 14 per cent in 1850; under £2,000 in most later periods). One important element of produce exports in the 1860s was the groundnut. The export value of this commodity rose from a mere £1,611 in 1840 to ten times that value in 1850 and to almost £50,000 (22 per cent of the total) in 1867. The decline was scarcely less dramatic: £14,403 in 1880, £11,016 in 1889 and a mere £501 (less than 0.1 per cent) in 1900. Having regard to twentieth century patterns of production it is tempting to suggest that this decline was a direct result of the political changes which interfered with an interior trade, but Alldridge's comment on this matter suggest that the crop  had been widely grown in the coastal areas before adverse price policies cut down its profitability in the 1880s.²

The interior trade at its height in the 1850s and 1860s may, if the assumptions of the last two paragraphs are correct, have amounted to between 30 per cent and 40 per cent of Freetown's produce export trade. The routes along which these goods were moved were investigated by Major Festing, Assistant Colonial Secretary and later Political Officer at Sierra Leone, initially on the basis of interviews with incoming traders and later, in the field, where he died. In 1887 his interpretation of this material was published in diagrammatic form ³ which, though not entirely accurate, is sufficiently consistent to enable mapping of the principal routes (Fig. 1). In the interior areas Siguiri and Faranah stand out as significant nodal points, whilst Timbo, Musaia, Bumban and Samaia were the principal inland centres where traders collected before striking out to the coast at the tide-water port of their choice--Kambia, Mange, Port Loko or Magbile.

Quite apart from local disturbances among the coastal people which had a limiting effect on the total level of trade, this interior

¹ Pollett, J. D. The Geology and Mineral Resources of Sierra Leone, H.M.S.O. 1952, p. 20.
² Alldridge, op. cit, p. 341.
³ Sierra Leone Despatch, African 332, Colonial Office 1887, p. 53.

trade suffered special disruptions in the 1870s and 1880s brought about by inter-tribal warfare and the

increasing activity of the French in diverting produce to their own centre of Conakry. The final break, though it was never quite absolute, came with the work of the Anglo-French Boundary Commission and the Protectorate Ordinance of 1896 which circumscribed the hinterland of the Colony, imposing a rigid outer limit which excluded much of what might be termed the port's "natural " supply area. Sherbro, where local produce was overwhelmingly predominant, suffered little as a result of the delimitations.

Accordingly, the problems of the newly established Colony and Protectorate were seen largely in terms of Freetown. An Administrator-in-Chief of the period, writing to the Colonial Secretary had "...under consideration the important question of stimulating the production of exportable articles in the Protectorate...The loss in export trade caused by Freetown ceasing to be a depot for the rivers and territory now under French administration can be more than compensated by the development of the Natural Resources of the Protectorate".¹ The basis of this development was to be the expansion of the oil-palm industry, the products of which were finding a ready market in the growing urban and industrial centres of Europe, especially in Britain and Germany. An increased export of palm kernels could only be achieved by a revolution in transport methods which would permit an extension of the geographical limits of exploitation. The Sierra Leone Government Railway was to provide the basis for this revolution.

It was largely due to the energetic approach of Governor Cardew that action was quickly taken to put these ideas into concrete proposals and push them through to fruition. Not only did he press for the railway scheme in Freetown, London and on Merseyside but, by his tours in the hinterland, made personal assessments of the levels and direction of trade then current. On the basis of the regional analysis which he was able to put forward, a suggestion of potentials was made and this in turn dictated the decision to carry the railway line to the south-east into the Mende-populated areas. It will be clear from figure 1 that this decision represented a radical re-orientation of the traditional trading patterns, although the new emphasis was not entirely original (Fig. 2, Route III). The historical

¹ S.L.A. Governor's Letter Book, 1898-1900, Local No. 405, p. 160.

geography of the railway has been fully examined elsewhere.¹ This note is concerned with the short period following the partial closure of the interior caravan routes but before the impact of the railway was widely felt.

In a most important speech to the Legislative Council in May, 1895, Cardew described the major trade routes of the Protectorate in the following terms:--

"(1) From FALABA by BAFODEYA or KOINADUGU through KAHRENA or BUMBAN thence to PORT LOKKOH
(2) From MATOTOKA, tapping the KUNIKI and the SANDA LOKKO countries (the latter in the region of KARRIYEMA) thence by road to BENKIA on the ROKELLE RIVER where the produce is loaded in small canoes and transported to MAGBILEH and thence transhipped in larger ones to Freetown.
(3) From MONGERI through SENEHUN to Freetown.
(4) From the UPPER MENDI Districts to MAFWEH, PUJEHUN and BANDASUMA respectively, thence by waterways to BONTHE, LAVANA, SULIMA and MANO SALIJA.

As to the produce conveyed along these routes: By the first route small quantities of ivory and gold

from the interior, rubber from the interior but also from KURANKO. Cattle in considerable numbers, Kola nuts, small quantities of palm kernels, and oil and rice chiefly from the districts around PORT LOKKO.

By the second route, Palm kernels, rubber, benni-seed, Kola nuts, rice etc.

By the third route, rubber and Kola nuts, a little ivory and a small quantity of palm kernels.

By the fourth route, rubber, small quantities of Kola and ivory and large amounts of Palm Kernels and Oil".²

As will be clear from a comparison of figures 1 and 2 the first of Cardew's routes was in the main a beheaded and simplified version of the traditional "interior" route; the diversion via Koinadugu being the major innovation.

Some produce--gold, ivory, and cattle--was still crossing the frontier. The cattle trade survived French taxation and in April,


¹Dalton, K. G. The Road and Rail Networks of Sierra Leone. Unpublished M.A. thesis, University of Leeds.
² S.L.A. Minute Book of Legislative Council 1895-1898, pp. 15-27 and p. 38, meetings of 1st and 3rd May, 1895.


1899, as many as 189 cattle traders passed through the Koinadugu District. Rubber also continued to be brought across, but at a much reduced rate; sheep, calabashes and "country" tobacco were minor imports. A District Commissioner reported on the situation; "The French are trying their utmost in every way to ruin the trade of this District which comes from and goes to French Territory almost entirely." Nevertheless, local production was still an important element in trade, and in some months of 1899 as many as 280 traders were collecting rubber, country cloths, pots, mats kola nuts and livestock within the District. The major articles of barter were cotton goods, salt, gin, beads, and matchets imported through Freetown and palm oil brought up from the coastal regions. In contrast to the large numbers of itinerant traders moving along this route (over 1,000 in April and 747 in October of 1899) there were only eleven settled traders in the upper Koinadugu centres and these were all concentrated in Falaba.¹

In the lower country served by these routes (Karene District) there were also numbers of itinerant traders, but the majority were located at the tide-water settlements, about a dozen in each major town. As well as the trade which came down the routeways, coastal produce--husk and clean rice, palm oil and kernels, benniseed, rubber and kola nuts were collected and exchanged for the usual variety of "European merchandise". In all some seventy traders were so engaged during 1899.

Freetown was the final outlet for this produce and so the quantities coming from the northern districts cannot be distinguished in the official data.

Fortunately this is not the case for the southern group of routes. The pre-eminence of this area in the production of Sierra Leone palm kernels has already been indicated. In the more general category of "produce", the proportion was lower, some 31 per cent as suggested by detailed but incomplete 1899 data. About 10 per cent of the benniseed but less than 0.1 per cent of the rubber exports of Sierra Leone came from this region in that year.² The movement of produce in the coastal area was greatly facilitated by the network

¹ District data for 1899 and 1900 are from a MS. volume Trade Report 1899, in Sierra Leone Archives, a contemporary compilation of Minute Papers (also in S.L.A.) submitted by District Commissioners.
² It is perhaps worth noting that the shipping Returns of the Blue Books greatly under-stress the importance of Sherbro, since most ships loading there also called at Freetown where they were officially "entered" and "cleared".


of waterways usable by small boats, but above the "fall line" towns of Mofwe, Pujehun and Bandajuma the streams were unnavigable and the effective trade area did not extend far into "upper Mende". Indeed it is clear from the discussions of the railway project that this was to open virtually "virgin" territory, certainly so far as heavy commodities were concerned. Significantly Bandajuma District failed to render the monthly trade returns called for in 1899 and Panguma district recorded no produce traders, though the score of traders in gin, tobacco, European cottons and provisions must have exploited some local purchasing power. In January, 1900, the Panguma District Commissioner was able to record an increase in the number of travelling traders in his district, but the limited upper Mende contacts with Freetown were mainly direct and overland rather than by the rivers and Bonthe.

Further hindrances to expansion in this area were the unsettled state of the country of which the Hut Tax war was only the final manifestation; and the physical problems of shipping produce--surf and shifting sand bars in the south, sand banks, silting of channels and the increased size of merchant vessels (1860--average ship--120 tons; 1900--1,200 tons¹) on the Sherbro river itself. Just as the role of Sherbro as a whole was by this time subservient to Freetown, so Bonthe dominated the other Sherbro ports. A sample of customs receipts in the latter part of 1884 shows Bonthe as having 92 per cent, Mano Salija 7 per cent, Sulima 0.3 per cent and Lavana a mere 0.1 per cent of total Sherbro trade.²  Later shipping returns, though difficult to interpret, suggest slightly larger shares for the three subsidiary ports, but Bonthe probably maintained well over 80 per cent of the trade.

The overland route from the upper Mende regions, produce being collected at Mongeri and transported through Senehun, ran transverse to the major drainage lines into the Colony proper and to Freetown (Fig. 2, Route III). As with the northern overland route only produce with a high value/weight ratio was important, though a few palm kernels were collected on the lower stages. By 1899 when more details became available, the railway was being extended along the general alignment of the lower portion of the route. Local trade in foodstuffs was greatly stimulated by the demand from the construction camps; some rice being brought inland from the coast about Shenge and Mando. Rubber and

¹ Means computed from Sherbro shipping data in Blue Books.
² S.L.A. Bundle 1883-1885, Sherbro

country cloths continued to come down from as far as Kailahun (officially in Liberia at this time) while from the middle Mende areas more rubber, benniseed, kola nuts, appreciable quantities of gum and "featheredge" boards were produced. At the height of the trading season of 1899 more than seventy five men and women had brought cotton goods, spirits, soap, beads, matchets, earthenware and kerosene into the Ronietta District for barter, and cash also was assuming importance as a medium of exchange.

Cardew's description of the second, the Rokel, route raises a number of problems, but also draws attention to its special characteristics. The problems arise from the outline of tributary areas, viz. "...tapping the KUNIKI and SANDA LOKKO countries (the latter in the region of KARRIYEMA)". Kuniki provides no problems, this is the Temne section of the plateau watersheds between the Pampana and Sewa basins (Fig. 2), immediately to the east of the collecting centre of Matotoka. But the Sanda Lokko territories lie in an entirely different area astride the northern routes. In figure 2, the whole Sanda and Lokko area is indicated and the chiefdom of  Sanda Lokko itself particularized. It seems extremely doubtful that trade from this area would ever use the Rokel river route, since it was well served by established overland route linking directly with the tide-water ports of the Great and Little Scarcies and more especially with Port Loko itself. Positive identification of Cardew's town of Karriyemma would solve the problem, but neither in contemporary literature nor maps has the name so far been found. There are two possible candidates: Karima, immediately to the east of the Lokko area and directly north of Benkia and the Rokel (Fig. 2), but it seems highly unlikely that goods would be collected on the eastern margins of a production zone for movement to the south-west. The second alternative identification appears much more likely: Kayima, a Kono town of the Sewa valley. It lies further to the east of the Kuniki country, in the Sando chiefdom, which borders on the Sanda Kuniki chiefdom. The ritualistically important Pampana River has the alternative name of Sanden. These toponomic similarities in an area naturally tributary to the Rokel trade route (Fig. 2) seem sufficient to account for a suggested slip of the tongue on Cardew's part. The fact that this was not challenged at the subsequent reading of the minutes is an interesting commentary on the geographical knowledge of the Protectorate's administrators at this early date.






The Rokel River route, extended eastward to its Kono tributary areas in the manner suggested, was an entirely logical outlet for the region it served. Although it was necessary higher up to cut across the major drainage lines, the plateau watershed, though forested, presented a more speedy journey than was possible through the marshes and forest along the unnavigable Sewa. The only partially navigable Pampana-Jong led southward to the coast to subsidiary Bonthe not westward to Freetown. Moreover, once Benkia on the Rokel was reached, the laborious and expensive human porterage was largely over and a much smaller group of men (requiring less food) could convey the goods to their final destination. Basic to the popularity of this route was the unique navigability of the middle reaches of the river, a direct consequence of the geological structure and physiographic history of the country.

The Rokel rises in the north-east of Sierra Leone and as it flows, initially south-westward, later more nearly due west, it crosses all the major rock groups which are banded roughly parallel to the coast. Apart from the narrow strip of the coastal Bullom sands, clays, grits and gravels, which are recent and relatively unresistant to erosion, the rock groups are old and also (Fig. 3), with one most important exception, hard; this is the Rokel River Series of pre-Cambrian sandstones and shales. This latter rock group, though by no means "soft", is appreciably less resistant than igneous and metamorphic rocks which surround and underly it.¹ Within this belt of sedimentary rocks, the Rokel River is seen to meander undisturbed by the rapids which impede its passage both in its upper and lower non-tidal reaches (Fig. 3). During the 70 miles of its middle course it falls little more than 70 feet, a direct contrast with the section immediately above Magbile where there is a greater fall in less than 10 miles over seven or eight sets of rapids.

But it would be a gross simplification to suggest that these contrasts in the stream profile result only from the relative erodibility of the rocks over which the river passes. The physiographic evolution of Sierra Leone is not yet fully  understood, but it is clear that, despite some movements in the opposite direction, the landscape has been moulded in large degree by forces related to a fall of sea level in relation to the height of the land. At each successive stage of this emergence, the rivers of the area have attempted to readjust their profiles to the new base level, cutting both downward and

¹ Pollett, op. cit.

headward into the country rock. This process naturally took place most slowly where the rocks were

hardest. The point at which headward cutting is currently taking place (technically, the "knick-point") may be marked by falls, rapids, or in soft rocks,by a local quickening of the rate of stream flow.

The rapids above Magbile mark the "knickpoint" of the current phase of down-cutting, and are mirrored in the rapids above Kambia, Mange, Port Loko, Matru, Mofwe, Pujehun and Bandajuma on the other major waterways of Sierra Leone which have the same origins. The broad middle reaches of the Rokel, mark a zone across which the knickpoint associated with an earlier higher sea level quickly migrated. It may now be traced in the rapid reaches of the river above Magburaka (Makump, on the opposite bank, is shown in Fig, 3).

Somewhat similar, but more restricted, middle reaches occur in the other streams crossing the Rokel River Series, but either because the volume of flow was insufficient, the lower rapid course too long, or the direction unhelpful, they have not come to be used for long-distance trade. With some portages at the major rapids and transshipment at Magbile it was possible on the Rokel alone to use energy-saving water transport for the movement of relatively bulky goods such as palm kernels and rice from inland zones a distance of over 130 miles to the major centre of Freetown.

Stress has been placed on the river route; there is some evidence that an overland route, parallel to the river, was also used as far down as Magbile.¹

There is no certain indication of the numbers of traders who were using this route, or were settled in the towns of Magbile and Rokel at the point of transhipment. In August, 1899, it was said that there were eight traders in the Marampa area, and Magbile was one of its chief ports, but this number, even at a slack period of the trading year so soon after the disturbances seems small in relation to the known size of these settlements.

In this discussion, two mechanisms for trade have been implied.

¹ S.L.A. Route Book, 1878-1888, p.83, includes the following information about the Yeeleh country (Yele is a town in Bonkolenken chiefdom 18 miles south of Matotoka): "...It is a distance of 3 days' walk to Magbellie one of the principal trading ports in the Marampa country in the north bank of the Rokelle (the latter opposite Magbellie on the Masimerah land) are also their trading ports in time of peace. These ports are also of same distance as Magbellie is to Teanneh." Tane is the chiefdom centred on Matotoka.


The first was the traditional caravan organized by the producers themselves or their neighbours; goods were collected in the interior for transport to the coastal centres where the exchange merchandise was stored in defensible factories or warehouses. Itinerant traders formed the second type of mechanism; whatever their origins these traders began each season with a stock of barter merchandise brought from the coast. During the season they travelled through the producing areas, selling or bartering their stocks and obtaining quantities of local produce which they themselves eventually took or sent to coastal traders and trading firms. No doubt there was some overlap of the two procedures, but the second must have been the more potent in stimulating trade and production in areas still following subsistence economies.

The mechanism of the "caravan" had characterized the earlier pattern of trade with the far interior, but was also operative within the restricted area of the protectorate. Water-borne trade of this type appears to have been typical in Sherbro;¹ it was also known on the Rokel. Some of the individuals who took part in these Rokel journeys still survive, and the opportunity was taken on a recent field visit to Matotoka of discussion with them. One informant's account was particularly clear.²

It was essentially a family undertaking; the informant's father made the trip four times before the railway was brought near enough to disturb the old balance of advantages. On all but the first trip the son accompanied his father and brothers. Although this particular family was new to the route in 1899, there was much local experience of it and at that date it had been travelled "time out of mind". The vessels used were dug-out canoes, manufactured by a "specialist" at the site in the forest where a suitable tree had been found and felled. On completion they were dragged through the forest by a large number of the retainers of the family to the nearest point on the Rokel.

November and December were the months for these journeys; the rains were only recently over and the rivers still carried enough water to float a canoe over the smaller rapids and sandbanks of the middle reaches. But river conditions were not the major factor; of greater import was the fact that this was a slack period in the

¹ Well described by Alldridge, op. cit., chap. II.
² I am greatly indebted to Paramount Chief Bai Kaffary, Pa Yamba Koré, Pa Gboro Kureh and other elders of the Tane Chiefdom for the material upon which the concluding section of this note is based.


farming calendar, and manpower might be spared from the village farms.

It was normal for several canoes to travel in company, each having a crew of six. At the major rapids it was necessary to unload the canoe and carry the cargo round the dangerous shallows where it was re-embarked after the lightened canoe had been floated through with a skeleton crew of two. Occasionally the canoe was upset, and the two "Captains" thrown out, but loss of life was rare. Usually six sets of rapids were sufficiently difficult to demand this laborious transhipment.

From a place (Manepe) a few miles above Benkia to Magbile normally took ten days, while the return journey, against the stream and with the canoe having to be manhandled up the rapids half as long again. Rapids were not the only hazard; at one point (Rokorbana) crocodiles abounded. As the vessels were loaded to the gunwhales the crews increased the freeboard (and their feelings of security) with a rough fence of inter-woven sticks which might be removed once the danger was passed.

At Magbile the cargoes were transferred to the much larger sprit-sailed cutters by which the journey to Freetown was completed in a day; each cutter could carry the cargo of ten canoes. Only two or three men from each canoe accompanied the goods on this final stage.

A stimulus to the family's interest in this route was the imposition of house tax, to pay which they wanted money. Tax might be paid in kind, but as it had to be delivered at Kwelu, five days journey, it was obviously preferable to carry cash rather than a miscellaneous load of rice, fowls, mats, goats, country cloths, kola and benniseed. On their first trip to Freetown they were unfortunate in that cowries only were proffered for their produce. On the second trip the canoe-load of thirty hampers of rice, benniseed and palm kernels, of which twenty-five were being carried for other families, yielded 300 cowries allowing the purchase of salt, tobacco, cloth, spirits, beads, small black skull caps and other merchandise, for later distribution in the village. On this occasion, the Magbile to Freetown section of the journey was made in a boat, the "Barifa" owned by one Pa Santigi Bangura.

Coins were obtained on each of the two succeeding journeys and, although there was some superstitious dread of them, a few were brought back to be shown to the villagers.

By 1908 the railway had reached Roruks and trade at the railhead replaced the laborious, but stimulating, Rokel River journey. An early trader at Roruks was the construction engineer, Mr. Harry Gilbert.¹ About the same time the collection centre for taxes was brought closer, first to Makeni and later to Mabonto. Therefore although the railway brought a new mode of transportation some of the earlier stimuli were removed.

This review of trade routes during a very short phase of Sierra Leone's economic development may serve as a contribution towards a historical geography of the country. Al;so, in stressing the first-hand account of the Rokel River route, one wishes to show the importance of eye-witness testimony for the reconstruction of the geography of a period not abundantly documented and which although within living memory, may well not be so in another decade.

¹ The name only was supplied by informants at Matotoka, but it seems likely that this is the same individual whom Alldridge met near Baiima in 1909. Alldridge T. J. A Transformed Colony, (Seeley, 1910), p. 171, et seq.