Price Le 4.00.

Bulletin Number 5


Volume 1 – Text

Volume 2 – Maps

Price Le 4.00

The diamond mining industry is one of the mainstays of Sierra Leone’s
economy. In 1968 diamonds accounted for 62% of the total value of exports. For
1969 the percentage will probably be higher, judging from the monthly purchases by
the Government diamond Office which buys nearly all of the diamonds produced
under the Alluvial diamond Mining Scheme.

Because of the prominence of diamond mining in Sierra Leone, this Bulletin,
which represents the outcome of seven years extensive fieldwork by the author, is of
major importance. In it Mr. Hall gives a comprehensive review of the history of
diamond mining in Sierra Leone and also of mining practice, in addition to a detail
account of the geology and geomorphology of the diamond fields.

It should be noted that the opinions expressed in this Bulletin are solely those
of the author and do not necessarily represent the views of the Director of Geological
Survey. This in particular applies to the estimates of the extent of smuggling and
illicit mining.

Geological Survey Division
Ministry of Lands, Mines & Labour
Freetown, Sierra Leone









(i) Location, extend and communications

(ii) Climate and vegetation

(iii) Topography



(i) The discovery of diamond

(ii) Exploration and development by Sierra Leone Selection Trust

(iii) Illicit mining, licensed mining and smuggling

(iv) Official investigations

(a) Diamond corporation Technical assistance projects

(b) The present survey



(i) Sierra Leone Selection Trust operations

(ii) Dredging by the Diamond Exploration Company

(iii) Licensed mining

(iv) Illicit mining

(v) Mining costs and minimum grades



(i) Prospecting by alluvial miners

(ii) Suggested methods



(i) Description of Sierra Leone diamonds

(ii) Diamond prices

(iii) Current production

(iv) Revenue from diamonds






(i) Rock formations

(ii) Structure


(i) Controlling factors

(ii) The drainage pattern

(iii) Drainage classification and stream order

(iv) River profiles

(v) River terraces

(vi) Swamps

(vii) Laterite

(viii)Planation surfaces

(ix) Age and correlation of surfaces



(i) Classification of diamond deposits

(ii) Types of alluvial deposit

(a) Summary

(b) River channel gravels

(c) Recent flats

(d) Low terrace gravels

(e) High terrace gravels

(f) Swamp gravels

(g) Penetration deposits

(iii) Mineralogy of alluvial deposits

(a) Summary

(b) Minerals with alluvial significance

(c) Minerals with genetic significance

(d) Other minerals








Koidu area
Yengema area
Lower Moinde
Tefea area
Upper Sewa
General remarks

Block 1 – Peyima
Block 2 – Middle bafi
Block 3 – Nimikoro



Block 4 – Nimiyema
Block 5 – Jagbwema
Block 6 – Barma
Block 7 – Boajibu
Block 8 – Kobundala
Block 9 – Baoma Oil Mill
Block 10 –Leveuma
Block 11 –Yamandu
Block 12 –Jomu
Block 13 –Sembehun
Block 14 –Wubunge
Block 15 –Petewoma
Block 16 –Hima
Block 17 –Sumbuya
Block 18 –Bagbo



Tongo drainage
Woa and tributaries
General remarks



Block 19 – Bundoye
Block 20 – Panguma
Block 21 – Foindu
Block 22 – Beeya
Block 23 – Kenja
Block 24 – Lower Woa
Block 25 – Male
Block 26 – Putehun
Block 27 – Segbwema
Block 28 – Middle Moa
Block 29 – Zimmi
Block 30 – Moro
Block 31 – Malema


Teye (Mongeri)
Koye (Selu)
Kendi (Gau)
Lower Waanje






(i) Dykes

(ii) Pipes

(a) No. 1 Pipe

(b) No. 2 Pipe

(c) Other Pipes

(iii) Dating the intrusives







(i) The concept of alluvial dispersion

(ii) Objections to the alluvial concept

(iii) Source-rock investigations

(a) SLST kimberlite investigations

(b) Diamond Exploration company kimberlite search

(c) Geological Survey investigations

(iv) Summary



(i) Kimberlite pipes

(a) Prospects of success

(b) Recommended areas

(c) Exoploration methods and costs

(ii) Other sources

(a) Suggested target areas

(b) Methods of search





(i) General remarks

(ii) Basis of estimates

(iii) Grade classification

(iv) Summary of production and resources data

(v) Analysis of production data

(vi) Analysis of diamond resources

(vii) List of important remaining alluvial deposits



(i) Future production from the known deposits

(a) Production by Sierra Leone Selection Trust

(b) Production by licensed miners

(c) Posisble production from deep alluvials

(ii) Augmenting the reserves

(a) New alluvial fields for licensed mining

(b) Underground resources





Detailed production and reserve estimates

Summary of block totals from production and reserve estimates

Recorded annual diamond exports, and estimates of smuggled caratage

(Volume 2)



1:1 million

Occurrence of diamonds in Sierra Leone
Key to 1:50 Diamond Series


Generalised bedrock geology
Key to other Bulletins on the region



Planation surfaces


Block boundaries
Areas of diamond occurrence
Kimberlite zones



Alluvial diamond deposits
Mining areas

Sheet numbers and titles correspond with the following

Sierra Leone 1:50,000 Topographical sheets:
Sheet 58 – Yengema with parts of Sheets 47,57 and 59
Sheet 68 – Fala
Sheet 80 – Boajibu
Sheet 81 – Panguma

Sheet 82 – Banumbu

Sheet 90 – Bo

Sheet 91 – Blama

Sheet 92 – Kenema

Sheet 93 – Daru

Sheet 100 –Sumbuya with part of Sheet 89


The Sierra Leone diamond fields cover an area of 7,700 square miles in the south-eastern part
of the country, and are centred on the town of Kenema, from which the Alluvial Diamond Mining
Scheme is administered. The first diamond was found in 1930, in Kono, and subsequent exploration
established that there were extensive alluvial diamond fields in many part of the region, and an
exceptionally rich group of deposits in the Koidu area.

Sierra Leone overlaps the margin of the West African craton, and the diamond fields are
situated almost entirely on early Pre-Cambrian basement formations within the craton. The principle
rock-type is gradodiorite gneiss, but this contains inclusions of many metamorphic rocks, notably
amphibolites, ultrabasic schists, and granulites. The granitic rocks are cut by several fracture-systems
which exercise close control over the drainage pattern, and by numerous narrow dolerite dykes.
Topographically the diamond fields consist of the undulating coastal plains, and the dissected margins
of the inaldn plateau, which are embayed by the broad valleys of the major rivers. Plateau surfaces,
displaying subdued topography similar to that of the plains, occur at various elevations between 750
feet and 1,500 feet, and form a step-like sequence. Much of the plateau has, however, been thoroughly
dissected, giving rise to rugged hilly country. There are, in addition, three outstanding ranges of steep
forested hills which make the position of major schist belts within the granite. Drainage density
averages four miles of permanent drainage channel per square mile, but the first and second order
channels are usually swamps, which in fact constitute 80% of the dense network. All rivers have a
stepped profile, with rock-bars every two or three miles; mean gradients of less than 4 feet per mile are
uncommon, and the major rivers are rarely able to form a wide flood-plain.

Alluvial diamond concentrations occur in river channel gravels, flood-plain gravels, terrace
gravels, gravel residues in soil, and swamps. Values in these deposits vary over a very wide range, but
most important of which is corodum. When waterworn, this indicates the presence of alluvial residues
in soil and swamps; in stream gravels it is useful as an indicator of points of alluvial concentration.
Magnesian ilmenite and pyrope are present in a few deposits, and indicate the proximity of kimberlites.

The distribution of alluvial diamonds reflects not only the distribution of source-rocks, but
also past and present drainage routes. The location of these controlling factors has been principally
determined by major lines of structural weakness in the granitic basement, which therefore exercise a
predominant influence over diamond distribution. The most important diamond fields are those of
Kono, Tongo, and the Sewa Valley. In the Kono and Tongo fields, the deposits are mostly in the flood-
plains and low terraces of streams and small rivers. In the Sewa Valley, however, the principal
deposits are those in the channel of the Sewa itself, a wide major river with a rocky irregular bed.
There are also narrow terrace deposits beside the river, and swamp deposits are scattered over the
valley floor. Groups of swamp, stream and terrace deposits occur in many other parts of south-eastern
Sierra Leone, but these are of lesser importance in terms of production.

Kimberlite outcrops have been discovered in two separate areas: the Koidu area, in Kono, and
the Tongo area, 30 miles to the south. Almost all the kimberlite bodies are in the unusual form of
narrow anastomosing dykes of uncontaminated, porpyhritic kimberlite. Five small pipes have been
found, which, although they consist principally of typical kimberlite breccia, also contain bodies of
inclusion-free kimberlite and breccias of abnormal texture. Underground development of kimberlites
has recently begun in the Koidu area.

The two kimberlite areas constitute only a very small part of the diamond fields and some
high-grade alluvial deposits are 70 miles from the nearest known kimberlite group. Nevertheless,
heavy mineral sampling campaigns have failed to detect significant amounts of kimberlitic indicator
minerals anywhere except in the vicinity of the known zones, and most workers have therefore
concluded that virtually all the diamonds in the fields have been derived by alluvial processes from the
Koidu or Tongo source areas. However, data recently assembled by the Geological Survey point to the
opposite conclusion, that is, that there are diamond sources in all parts of the fields. The nature of these
sources remains unknown, but they are seemingly deficient in the normal indicator minerals, and they
may well differ from kimberlites in other important respects.

– 1 –
Sierra Leone Selection Trust began alluvial diamond mining operations in 1932, and has now
produced a total of 19.6 million carats from the Kono and Tongo deposits. Over the last ten years,
production has averaged 650,000 carats per annum. * The company’s alluvial reserves appear to be
adequate to sustain production at present levels for at least twelve years; whether any part of the
diamond resources in kimberlite bodies can also be classified as payable reserves is not yet established.

Illicit mining and smuggling of diamonds began in 1950; the illicit mining was superseded by
licencsed mining in 1956, but it is only in the last two years that smuggling has finally been reduced to
negligible proportions. Alluvial diamond mining by licensed diggers has now become an established
and stable feature of the economy, and provides income and employment for about 25,000 men. The
total past production from licensed and illicit mining is estimated at 12.8 million carats, of which 4.9
million carats are estimated to have been illegally exported. Annual production is now about 800,000
carats per annum.

Total resources of alluvial diamonds remaining in the known fields, excluding SLST reserves,
are estimated to be 16.6 million carats, but 6.7 million of these are in deep gravels whose exploitation
will entail heavy capital investment and skilled management. It is considered that several small alluvial
fields remain to be found, and a systematic search for these will soon be necessary, if a steady decline
in licensed mining is to be averted.

Important underground diamond resources are almost certainly present in kimberlites or other
source rocks in many parts of the established alluvial fields. Some difficult in finding these source
rocks is anticipated, because of the apparent lack of indicator minerals, but a list has been compiled of
suggested target areas. Finally, exploration of certain sections of the country for kimberlite pipes is
most strongly advocated, even where there are no extensive alluvial diamond fields.

*these statements relate to 1965.

– 2 –


The alluvial diamond fields of Sierra Leone have produced over 32 million carats of diamond
since they were discovered in 1930, and there are still substantial undeveloped resources, both alluvial
and primary. In this Bulletin, the geology of the diamond deposits is described, and an attempt is made
to assess the resources and the future of the fields. In order that a completed picture should be
presented, a brief account has also been given of the historical, practical and economic aspects of
diamond mining in these fields.

Reference to various alluvial deposits which are believed to contain substantial unmined
reserves will be found amongst the descriptive matter of Part III. The more important of these have
been listed in Part V, Section A (vii), and this list is commended to the attention of interested parties.
It should be borne in mind, however, that the alluvial mining situation is far from static, and so in the
interval between completion of the field work and publication of this Bulletin some of the
recommended deposits may have been exploited. Readers who are interested in these deposits may
justifiably complain that concrete sample results are not available. This is inevitable. If they could be
sampled with the limited equipment available to the Geological Survey, they would long ago have been
mined by the licensed miners. Most of them are under deep overburden or deep water.

The author wishes to express his grateful thanks to the many thousands of people without
whose assistance the survey of the diamond fields could not have been carried out.

Field work has been continuously assisted by officers of the Mines Division, including the
assistant chief Inspector o9f Mines at Kenema, several Inspectors of Mines and Area Superintendents,
and numerous Wardens. Without the help of the Area Superintendents and Wardens, the location of
mining areas would have taken twice as long, and the voluntary co-operation of the miners would
frequently have been difficult to obtain. Thanks are due to the chiefs and members of the local
administration in all parts of the Bo, Kenema, Kailahun and Kono districts who have been helpful
throughout the survey, and also to the several thousand licensed miners who have provided gravel for
sampling, diamonds for examination, and stories for taking a pinch of salt with.

The author is particularly grateful both to the General manager, Sierra Leone Selection Trust,
for making mine plans and records freely available, and to the Company’s Prospecting Engineers and
Geologists, who have supplied information, assistance and controversial discussion on many occasions.

Four Prospectors of the Geological Survey have assisted in the work of the diamond survey:
B.G. Burdett (July, 1961 to February, 1962); D.K. Bongay (October, 1963 to February, 1966); E.R.
Gregory (June, 1964 to September, 1965); and M.N. Perrins (October, 1965 to February 1966). All
these officers were engaged principally on kimberlite investigations, and their meticulous work has
been much appreciated.



The alluvial diamond mining fields cover most of the Eastern Province of Sierra Leone, and
the Eastern half of the Southern Province. If a boundary is drawn to encompass all the fields, the total
area enclosed is about 7,700 square miles, but within this area only about 12% can be regarded as
diamond mining country, and the aggregate area of actual diamond-bearing alluvial ground is only 80
square miles, or about 1%. The town of Kenema, which is roughly at the centre of the diamond fields,
is the administrative centre of the Alluvial Diamond Mining Scheme, and is also the headquarters of
the Eastern Province.

– 3 –
Kenema and many of the other principal towns are linked directly to the capital, Freetown, by the
railway, which bisects the fields from east to west. The southern and central fields have a good
network of main and secondary roads, and are connected by main road to Freetown. The distance from
Freetown to kenema by road is 232 miles, much of it tarmac and the journey generally takes about six
hours. The northern fields, that is, those of Kono District, have an adequate road network, although the
surfaces are sometimes rather poor. The distance from Freetown to Yengema, via the new Masingbi
road, is about 200 miles.

There are airfields at Bo, Kenema, and Yengema, to all of which schedules services are run by
Sierra Leone airways.


The climate of the diamond fields is wet tropical monsoon, with a single wet season each year.
The average annual rainfall is about 100 inches overall, but it is generally a little higher in the south-
east and a little lower in the north. The greater part of this rain falls in the west season, from mid-May
to mid-November. The wettest month is usually august, but rivers attain maximum discharge in mid-
September. There is very little rain in December, January, and February. River discharge is at its
lowest in March and April, and begins to increase in May, but ground-water levels do not rise
significantly until late July.

The normal temperature range is 70° to 92° although it can exceptionally drop as low as 50°f
at night in Kono, during January. Day temperatures average 88°F in the dry season and 82° in the wet

The original vegetation throughout the diamond fields was tropical rain forest, but over most
of the region, the forest has been destroyed to make way for farms. Residual areas of primary forest
remain where the population is sparse, that is , in the Gola forest in the extreme south-east, in the
rugged country adjacent to the Sewa above Jaiama, and in the three major hill ranges. Elsewhere, the
land has all been farmed on the shifting cultivation system and is normally covered with dense
secondary bush whose height varies with the number of years that have elapsed since the ground was
farmed. In the north, where rainfall is less, there are large areas where the secondary bush has been
replaced by elephant grass, with shrubs and trees persisting only along the watercourses.


The coastal plains from three-fifths of south-east Sierra Leone, the rest of the regions
consisting of the embayed and partially dissected southern margin of the interior plateau. Most of the
actual diamond fields display subdued topography of low convex hills separated by a network of
shallow swamp and stream valleys, with occasional hills of monadnock type or bare granite inselbergs.
This topography is typical, not only of the plains of the south, but also of the valley floors between
plateau outliers, and of the plateau surfaces where they remain undissected.

Much of the plateau in this region has in fact been deeply dissected to form rugged country of
steep granite hills separated by broad stream valleys: where this has happened remnants of the plateau
surface may occasionally survive as flat hilltops. Although such dissected topography forms about
one-fifth of south-east Sierra Leone, diamond fields are rarely associated with it.

The outstanding physical features of the region are the three principal hill ranges, formed by
steep, forested ridges which rise 600 feet to 1,200 feet above the surrounding country. These are the
Kambui, Nimini, and Gore hills, and the backbone of each range is a belt of metamorphic rocks of the
Kambui Schist Series, striking roughly north-south.

– 4 –
The plains and plateaux represent, of course, a series of planation surfaces initiated during
difference periods of geological time; these surfaces are defined and discussed in Part II, Section B.

Part 1, Section C – HISTORY


The early history of the diamond fields has been described in some detail by Pollett (1937),
and much of the information given below has been abstracted from his account.

In January, 1930, a Geological Survey field party, consisting of the director, N.R. Junner, and
his Assistant Geologist, J.D. Pollett, was traversing the Kono District. Pollett’s traverse took him
across the Gbobora stream near the village of Fotingaia, and on examining the stream-bed gravels for
heavy minerals he recovered a crystal which was subsequently identified as a diamond. On the
following day, Junner recovered another diamond from the same site.

The discovery was reported in the Annual Report of the Geological and Mines Department for
the year 1929, but no interest was shown in it until Junner brought it to the attention of Consolidated
African Selection in the Gold Coast. A prospecting party from this Company arrived in Sierra Leone in
March 1931. In the same month, the first hint of the widespread nature of the diamond occurrences
was received when Pollett found two more diamonds in the gravels of the Kenja stream at Pava, about
51 miles south of the original discovery.


The CAST prospecting party began work at the Gbogora discovery site, but the first line of
pits across the flat produced only nine diamonds, weighing altogether 0.35 carats, from 25 cubic yards
of gravel. Other pit lines produced similar poor results. However, the occurrence of diamond had been
confirmed, and the Company applied for a Special Exclusive Prospecting Licence to cover the areas of
both the Geological Survey discoveries. Further prospecting in the vicinities of both discovery sites
continued for some time to produce poor results, but the Kono district appeared to offer greater
possibilities, so work was finally concentrated there. As the prospecting parties spread eastwards,
diamonds ere found in the Bandafaiyi and the Oyie, but it was not until mid-1932 that the first
unquestionably payable deposits were found along the Shongbo stream. Mining of the Shongbo
deposits began in the late 1932.

During the years 1932 to 1935, a rapid expansion of prospecting activity took place, and it
soon became apparent that the principal deposits of the field were to be found within an area of about

90 square miles, centred roughly on the village of Yengema. The depsoits near Koidu, probably the
richest diamond alluvials in the world, were found in January 1935, when three small prospect
pits,sited at the Woyie-Wongoyie confluence, produced 749 diamonds eighing 371 carats.

Early prospecting results had shown that all the Kono diamond deposits were located in
tributaries of the Bafi river, itself a major tributary of the Riva Sewa. A reconnaissance of this drainage
system was carried out during 1933, and it was found that good diamond values occurred not only in
the Bafi channel, but also right down the Sewa channel as far as Sumbuya, twenty-nine miles from the
coast. Subsequently, more detail prospecting estsablished the presence of the Tefea deposits near the
Bafi-Bagbe confluence, and showed that diamonds were present in many Sewa tributaries and terraces.

Sierra Leone Selection Trust was formed in 1934 as a private company, wholly owned by
Consolidated African Selection Trust. In 1935, an agreement was concluded whereby S.L.S.T. was
given the sole right to prospect for and mine diamonds in Sierra Leone for a period of 99 years. The
continuity of operations having thus been assured, a permanent mining camp was built near the village
of Yengema, and for some years, the resources of the Company were concentrated upon production
from the alluvials of Yengema Field.

– 5 –
Prospecting activities were resumed in 1946, and attention was initially focussed on finding
the source of the diamonds. Finally, in 1948, diamond-bearing bedrock, in the form of brown clay
dykes cutting the decomposed granite, was exposed in pits near Koidu and on the Oyie stream. The
dyke rocks were subsequently proved to be tru kimberlites and to form part of an extensive kimberlite

Between 1948 and 1955, exploration was carried out in many parts of Sierra Leone. Along
the Sewa River, prospecting parties delineated payable gravels in many of the Sewa tributaries and
terraces, and work began on the construction of mobile washing plants suitable for exploiting these
scattered deposits.

In the vicinity of Pottett’s Pava discovery, early work had shown that there were numerous
scattered deposits to the east and north-east of Kenema, but it was generally thought that they were
associated with the Moa River, and were derived from a source area outside Sierra Leone. In 1953, a
decisive prospecting campaign began in the Kenema district. Good values were found to be present in
the channel fo the Male River, a Moa tributary whose drainage basin lay entirely within Sierra Leone.
The diamond trail was followed up the Male river to the Woa, then up the Woa River to the Tongo, and
investigation of the Tongo stream led rapidly to the discovery of the high-grade alluvials of the middle
Tongo and lower Lando. The presence of indicator minerals in the Lando gravels pointed to the
existence of a new kimberlite zone, unconnected with Koidu zone. Construction of a mining camp in
the Tongo area began in 1955, and mining operations were started in 1956.

Elsewhere in Sierra Leone, reconnaissance prospecting parties visited all districts and tested
all the major rivers and their principal tributaries, washing bulk gravel samples from rock-bars. As a
result of this work, isolated diamond finds were made in several rivers, but subsequent investigation of
each occurrence failed to discover any associated diamond fields. The conclusion was finally drawn
that the only diamond fields of consequence in Sierra Leone were those which the Company had
outlined in the south-east part of the country, and this conclusion has not yet been invalidated.

Concurrently with this widespread exploration activity by the Company, the rapid spread of
illicit mining during the period 1950 to 1955 was creating an increasingly difficult security situation.
In 1955, the Agreement with the Sierra Leone Government was revised, and S.L.S.T. surrendered its
rights to the greater part of Sierra Leone. Under the terms of the new Agreement, alluvial prospecting
and mining rights were henceforth restricted to two defined areas at Yengema and Tongo, and were in
any case to terminate in 1985. This brought an end to S.L.S.T. alluvial explorations outside the
Yenegma and Tongo Lease areas.

The years 1956 to 1965 saw the steady expansion of production from the Tongo deposits,
augmenting the diminished Yengema production, and for some years now total production has
remained fairly steady, in the region of 650,000 carats per annum. One of the most interesting projects
of recent years has been the detailed investigation of the kimberlite development headings were driven
into one of the small kimberlite pipes in 1965.


It is by no means certain exactly when illicit mining of alluvial diamonds began in Sierra
Leone. Van der Laan (1965) suggests 1952, but discussions with individual miners and dealers
indicate 1950 as the more probable date. All reports agree that really serious illicit digging only began
in 1952, stimulated and encouraged by the arrival of numerous foreign miners and dealers, principally
from Guinea. In 1955, there were tens of thousands of men engaged in illicit mining, which by this
time had spread throughout the diamond fields.

– 6 –
The illicit mining chiefly took the form of pitting in swamps and shallow stream flats, although
deposits on rock-bars in river channels were also mined, and there was some diving in the Bafi and the
Sewa. There was eventually little attempt at concealment, since most of the deposits were not near to
the roads, and the strength of the Police and the S.L.S.T. Security Force was in any case inadequate to
repress the large numbers of determined diggers.

By 1955 the depletion of diamond resources and the breakdown of law and order in the diamond
fields had become serious problems, and the Government decided to seek a revision of the Agreement
with S.L.S.T. in order to be able to license and control the digging, and to divert the diamonds into
legitimate taxable channels. A new Agreement was concluded in October 1955, whereby S.L.S.T.
relinquished its rights to all the alluvial diamond deposits outside the Yengema and Tongo Fields in
return for compensation of £1,570.000.

In February 1956, the Alluvial Diamond Mining Scheme was inaugurated, and the issue of
licences to miners and dealers began. The scheme was immediately successful in bringing the digging
under control; by the end of March, fifteen hundred mining licences had been issued, and within a few
months, virtually all illicit diggings had been converted to licensed mining areas. Since that time, there
have been only rare instances of illicit mining in the alluvial diamond mining areas, although there
have been several serious outbreaks within the S.L.S.T. Mining Lease at Yengema.

Control of the diamonds produced by the alluvial miners was less easily achieved. Before 1956,
of course, diamonds could not legally be bought or sold in, or exported from, Sierra Leone, and all
illicit diamond production was therefore smuggled, i.e. illegally exported. From February 1956, the
marketing of Sierra Leone diamonds was entrusted to the diamond Corporation, whose buying
organisation, it was expected, would drive the smugglers out of business. Offices of the Diamond
Corporation (Sierra Leone) Limited were set up in Bo, Kenema, Boajibu and Sefadu, and in some areas
valuers travelled through the bush to mining sites. In spite of this, however, only a fraction of the
substantially increased alluvial production was sold to the D.C. S.L.I. in 1956, and most of the
diamonds continued to pass along the well-established smuggling channels to the outside market. *
Between 1957 and 1959 there was some improvement, and the proportion of the diamonds smuggled
dropped to less than half of the total production in terms of cartage; nevertheless, this still represented
over two-thirds of the total value. The pattern of diamond disposal that grew up was that the coated
stones and the small or poor quality clear stones were sold to the D.C.S.L., since these were not in
demand on the outside market, where the prices paid were higher than those offered by the diamond

The whole picture was radically changed in august 1959 with the opening of the government
diamond Office, which replaced the D.C.S.L. as the sole legal exported of diamonds (apart from Sierra
Leone Selection Trust, at that time). The G.D.O., which was managed by the Diamond Corporation on
behalf of government, was authorised to purchase gem diamonds for prices comparable with those
ruling in Monrovia, and this resulted in a dramatic increase in legal diamond exports. Since 1960,
smuggling has continued progressively to diminish in importance, and is now believed to be of
negligible proportions.

The picture now, ten years after the inception of the Alluvial diamond Mining Scheme, is one
of stability and predominantly legitimate operations. Alluvial diamond mining has become an
established and accepted feature of the economy of the country, and provides a living for about 25,000
miners, now mostly Sierra Leoneans. Smuggling is no longer a serious problem, and the miners now
receive a reasonable proportion of the export value of their diamonds.

* According to Van Der Laan (1965, this was centered mainly on Monrovia after 1955.

– 7 –

(a) Diamond Corporation Technical Assistance projects

In 1960, the Diamond Corporation undertook to provide the Sierra Leone Government with
technical assistance in the form of a number of mineral investigation projects. In addition to the
investigation of platinum and gold occurrences, these projects included:-

1. The assessment of the economic potential of some diamondiferous swamps which had
been abandoned after incomplete mining.

2. An investigation of the feasibility of mining the Sewa channel gravels by dredging.

3. A systematic search for bodies of diamondiferous kimberlite.

For the purpose of carrying out this work on behalf of Government, the Sierra Leone State
Development company was formed in 1960; in 1962, its title was changed to the diamond Exploration
Company (Sierra Leone) Limited, generally abbreviated to DEXSIL. Field work on the projects began
in January 1961, and ended in January 1965, and the results and conclusions were presented to
Government in a series of reports (Barber, 1961; Barber, 1963; Stracke, 1963; Forristal, 1965). A great
deal of invaluable data are contained in these reports, and the more important facts and figures have
been summarised in the relevant parts of this Bulletin.

Project 1 was the first to be undertaken, and three formerly-mined swamps were investigated
during 1961 and 1962: Sembehun 14 (Beeya, block 22), Sembehun 17 (Bakwe block 13), and Boama
(Papayei, block 9). The results from all these swamps clearly demonstrated that considerable wastage
of alluvial diamond resources was inherent in the unsystematic mining methods which had been used.
Although these swamps, in common with numerous others throughout the fields, are regarded as
worked-out, substantial reserves remain in gravel pillars, tailings and overburden, but it would now be
uneconomic to recover these diamonds by any mining method.

Project 2 began in 1962 with the construction of a specially-designed suction dredge at Boajibu,
on the Sewa River. Experimental dredging operations were carried out along a three-mile section of
the Sewa channel until January 1965, when the project was closed down. This work established that
the Sewa channel gravels could be effectively mined by suction dredging. The minimum grade for
profitable mining is about 0.2 carats per cubic yard, but the grade of the deposits within the area tested
was unfortunately far below this.

Project 3 was carried out from mid-1961 to mid-1963. The method of search employed was the
systematic collection and laboratory examination of heavy mineral concentrates from swamps and
streams throughout south-east Sierra Leone. Six areas, where apparently anomalous concentrations of
kimberlitic indicator minerals were found, were investigated in detail. No diamondiferous kimberlites
were found, and the DEXSIL geologists concluded that kimberlites of economic interest exist only in
the Koidu and Tongo source areas. This conclusion is not accepted by the present author.

(b) The present survey

In the situation that existed from 1935 to 1955, during which time Sierra Leone Selection Trust
had exclusive rights to Sierra Leone diamonds, the Geological Survey had little to do with diamond
deposits, and it slender resources were applied to the search for other minerals. With the ending of
these rights and the beginning of the Alluvial diamond Mining Scheme, the Geological Survey was
called upon from time to time to provide information, to prospect, or to assess deposits, and the
difficulties experienced at this stage drew attention to the extreme paucity of available facts on the
nature and extent of these important fields.

– 8 –
In November 1958, the author was appointed specifically to carry out a detailed survey of the
alluvial diamond fields, to investigate the geology of the deposits, and to assess the diamond resources.
The relevant data can readily be collected whilst mining is in progress, but are tedious and expensive to
obtain once the mining areas are abandoned, flooded and overgrown.

The examination and sampling of deposits proceeded for seven years, interrupted occasionally
by kimberlite investigations or prospecting campaigns. The length of time taken by the survey has
been due in part to these diversions, but principally to the very large number of individual deposits,
each of which requires one or more days for examination and sampling. Although all known diamond
deposits have now been examined, the work is not complete; in fact, a survey of this type can never be
complete, since new explosures yielding fresh data are constantly being made in the course of mining
in all parts of the fields. However, field work must be stopped at some point and the current state of
knowledge recorded as a basis for future work. That has now been done, and this Bulletin represents a
compilation of all the more important facts, figures, conclusions and hypotheses known to the author at
the beginning of 1966.



Sierra Leone Selection Trust carried out alluvial diamond mining in two separate areas,
Yengema field and Tongo field, but Yengema is the larger and more important field, and the mine
headquarters are there.

Valuation of deposits is undertaken by means of regularly-spaced lines of uniform six-foot
diameter pits. The gravel from the pits is placed in locked containers which are carried on tractor-
hauled sledges to a road, and then on lorries to a central washing plant. There, the gravel is treated in
vibrating screens and Pleetz jigs, and the diamonds are finally hand-picked from the concentrate.

Standard opencast methods are employed to mine the deposits. After damming and diversion of
the stream, overburden and pay gravel are stripped by dragline. Where the bedrock is hard and
irregular, it is sometimes necessary for gravel pockets to be scraped out by pick and shovel. The mined
gravel is carried in dumper trucks to one of several Pan Plants, where the gravel is screened, washing,
sized and jigged. The Pan Plant product is concentrate which contains all the heavy minerals from the
gravel; this concentrate is taken in sealed canisters to the central Concentrator House where it is again
sized before being passed over vanners and grease tables to collect the diamonds.

The Company operated a small experimental jet-lift dredge in the Bafi river from 1962 to 1964,
but recoveries were too low to cover the working costs.

A Contract Mining Scheme is run by the Company; under this scheme, groups of local miners
are encouraged to mine selected shallow deposits within the Company’s Mining Leases. The deposits
are those which contain proved payable ground but which would be difficult for the Company to mine
because of isolation or small size. Equipment and advice are supplied by the Company, and the mining
method consists of the manual extraction of gravel from successive parallel working cuts across the
flat, after diversion of the stream into a leat. All the working cuts connect with a single drainage cut,
placed generally along the side of the flat, and discharging lower down the valley. After extraction the
gravel is deslimed and sized in rockers (foot-shakers), and then jigged, first in a Joplin jig, and
subsequently in a small hand-jig (gravitator). The final concentrate is hand-picked, and all diamonds
won are purchased by the Company from the miners.


The Diamond Exploration Company (Sierra Leone) Limited, was a subsidiary of the Diamond
Corporation, and was formed to carry out certain specific investigations into diamond mining and
occurrence on behalf of the Sierra Leone Government.

– 9 –
One of the projects was to investigate the feasibility of exploiting the Sewa channel deposits by
dredging. For this purpose, a small self-contained suction dredge with a concentrate recovery plant on
board was designed by a consultant engineer, and was built in Sierra Leone in 1962. the three-mile
section of the Sewa just north of Boajibu was selected for the dredging experiment; this section was
easily accessible by road and was considered to represent a good cross-section of river conditions.

The dredge consisted of a barge formed of welded steel pontoons. On it was mounted a diesel-
driven centrifugal gravel pump, with a 12” diameter flexible suction hose. The pump discharged to a
pair of vibrating screens, and the screen products were fed to jigs and a picking belt. Jig concentrates
were taken to shore for re-concentration and hand-picking. Under good operating conditions, with
thick continuous deposits, a gravel pumping rate of 25 cubic yards per hour could be easily maintained.
After some spot prospecting, dredging began in May 1963 at the downstream boundary of the reserved
area, with the intention of moving slowly upstream dredging a series of parallel contiguous cuts across
the river.

Initial results showed that much of the gravel was unpayable and that to dredge the entire
reserved area would take at least twenty years. The programme of continuous dredging was therefore
dropped in favour of a series of sample cuts, each thirty to forty feet wide, and spaced at intervals along
the river. Dredging continued until January 1965, by which time the dredge had produced a total of
3,238 carats from some 90,000 cubic yards of sand and gravel, but had discovered no deposits which
could be regarded as payable.

The total expenditure on this dredging project was Le.624,256, and the net cost, after the sale of
diamonds and equipment, about Le. 500,000. A great deal of useful information was obtained, both on
the nature of the channel deposits and on the feasibility of suction dredging on the Sewa, and this was
presented to Government in a detailed report (Forristal, 1965) from which most of the forego9ing notes
have been abstracted. Notes on the river channel deposits will be found in Part III, in the description of
the alluvials of Block 7.

The section dredge amply demonstrated its ability to dig gravel and clean bedrock at depths from

5 feet to 60 fett, and it was considered that depths of up to 100 feet should be attainable. The pump
could handle material up to 8 inches in diameter, and blockages due to boulders, sticks or vegetable
debris accounted for only 3% of working time. The feasibility of suction dredging in the Sewa has
therefore been established, provided large yardages of adequate grade are present. From the data given
in the report, it would appear that the minimum grate for profitable operation is about 0-.2 carats per
cubic yard.


The Alluvial diamond Mining Scheme came into operation at the beginning of 1956, and has
been a most important feature of the economy, since, in most years, about 30,000 men have been
engaged in licensed diamond mining. A comprehensive account of the administration of the scheme
and of the mining methods employed has been given by Fiarbairn 1965; only the salient points will be
outlined here.

The Scheme is administered and controlled by the Mines division of the Ministry of Lands,
Mines and Labour, and is under the general supervision of an Assistant chief Inspector of Mines, who
is assisted by two or three Inspectors of Mines. Area Superintendents are responsible for the issue of
licences and are assisted by about 95 Wardens, who duties are to provide close field control of the
mining areas. Alluvial diamond Mining licences, which cover an area of not more than 400 square
feet, and which are valid for six months, can only be issued to Sierra Leone citizens or to Native
Companies. Current licence fees are Le. 30 for swamp, flat, or terrace ground, and Le.60 for a section
of river channel; one licence entitles the holder to employ up to 20 men. Dealers’ licences are granted
by the Chief Inspector of Mines, and these entitle the holder to buy diamonds from Mining Licence-
holders, and to sell diamonds either to other licensed dealers or to the Government Diamond Office.

The mothers employed by the licensed miners bear little resemblance to standard alluvial mining
practice. Prospecting of flat or swamp deposits usually consist of random pitting in search of
corundum. Valuation, as it is normally understood, is never carried out, although it could be done at
minimum cost under a Prospecting Licence, which costs only Le.10. In the channel of the Sewa River,
prospecting usually takes the form of locating a pool which has not been previously mined and whose
size, depth, and situation make it feasible to tackle. Sometimes, divers may go down to sample the
upper deposits of the pool, but the results so obtained are of limited significance. Along the Sewa,
however, the difficulty of obtaining samples prior to mining is of little consequence, as all pool
deposits contain diamonds, and at least half, payable gravel is present.

Three methods of mining the river-bed deposits are employed; diving, dam-building, and gravel
pumping by airlift. Skin diving has been popular from the earliest days of river mining, and is carried
out in relatively shallow water with the aid of a stick tripod, erected on the river bed. Although diving
was arduous and sometimes hazardous, very large numbers of men were formerly engaged in it every
dry season, and their enthusiasm was due to the fact that the bed of the river contained many small
pockets of abnormally enriched travel. Although a diver might bring up only ¼ cubic yard in the
course of a day, this was well worth-while when recoveries were usually between 2 and 15 carats per
cubic yard. In recent years, diving activity has considerably diminished, and many of the remaining
divers now use aqualung equipment.

The construction of earth-filled coffer dams, to seal off portions of the river bed, began in 1958,
and has been responsible for a major part of Sewa diamond production since 1960. Groups of licence-
holders often join together to form a Native Company for dam construction. The area enclosed by the
dam is pumped dry, and the gravel thus exposed is removed manually; as dam-building, pumping and
gravel extraction can take place only at times of minimum river discharge, all the work must be
completed within a period of about three months. Mined gravel is stockpiled, and treatment does not
begin until mining is completed, or is terminated by dam failure due to increasing river discharge in
early May.

Gravel pumping by pontoon-mounted airlift is undertaken principally by small Native
Companies, working along the Sewa and the Bafi. The equipment will only operate efficiently in
moderately deep water, and it has been successfully used to depths of up to 60 feet along the Sewa. It
has been noticeable that the most successful airlift operations are always those where aqualung divers
are used to locate gravel and position the pipe, whilst equipment which is unguided at the river bed
usually fails to bring up payable gravel.

Mining of swamps, flats and low terrace deposits is carried out exclusively by hand methods,
although small petrol-driven pumps are widely used for dewatering the pits. The work is done by small
gangs of tributors, who are not paid wages, but who receive a proportion, usually two-thirds, of the sale
proceeds of diamonds recovered. The licence holder has very little control over his tributors, and each
gang digs its own individual pits down to the gravel. Mining of low and medium grade deposits is
extremely selective, and the miners have developed considerable skill in removing only the gravel with
good values, with the most of the deposits are inefficiently mined. The mining is termed inefficient not
only because diamonds are left in the ground but also because the number of man-hours required to
extract and treat each cubic yard of gravel is unnecessarily high. The proportion of gravel left as pillars
depends on the depth of overburden, the values present, and the pattern of diamond distribution;
generally, the proporation is between 30% and 60% of the original volume. The resulting waste of
diamonds is not, however, so serious as would at first appear; for example, where 30% of the gravel
remains, it probably contains only 5% of the original values. High-grade swamps are usually
repeatedly mined until nearly complete gravel extraction is achieved. In the marginal and low-grade
deposits, net diamond recovery may by only 40% of the gross reserves, but it should always be
remembered that, because of the low values, these deposits could not have been mined at all by a
mining company using conventional methods. In fact, the estimated wastage of reserves due to
incomplete mining of payable deposits is roughly balanced by the estimated production attributable to
selective mining of unpayable deposits.

– 11 –
Serious wastage of resources only occurs in the deeper low terrace and river flat deposits, where the net
recovery from a gravel bed, containing 0.5 carats per cubic yard and overlain by 15 feet of overburden,
would typically be about 40% of the total diamond content.

High terrace deposits are mined by the same system of individual pits, but, ground water being
generally absent, pits in payable ground expand and coalesce and almost complete extraction of
diamond-bearing ground results.

The method of gravel treatment employed by the miners is the most unsatisfactory feature of
their operations, since, without exception, it results in incomplete recovery from the gravel extracted.
The same procedure is adopted for all types of gravel. The gravel is first washed by hand in perforated
headpans to removed fines and oversize material, and at this stage there are heavy losses of small
diamonds through the perforations, which vary from 2mm to 4mm in diameter. The washed gravel is
jigged in a small rectangular hand jib (known as a “kick”) with a 1mm wire screen, in order to
concentrate the heavy minerals. The jig is finally inverted on a level surface; the concentrate patch is
inspected closely for diamonds but the bulk of the jig product is only perfunctorily sorted through.
Whilst the jig itself is a reasonably efficient instrument, and is normally skilfully operated, the fact that
the gravel is unsized means that the concentration process is impeded, and the examination of the
jigged gravel is then too cursory to detect many of the diamonds, which have failed to gravitate.
Further losses therefore occur at this stage, chiefly of stones less than ½ carat, but also of occasional
larger stones which are flat or lacking in lustre.

Losses both in washing and jigging can be largely eliminated by using foot-shakers to wash and
size the gravel, but in spite of continuous pressure, advice, and assistance from the Mines Division, few
licence-holders can be persuaded to do this. Their reluctance is due, to a large degree, to the opposition
of the tributors, who find manual methods advantageous.

In the course of the alluvial diamond survey, an attempt has been made by the Geological
Survey to measure average losses in the miners’ gravel treatment operations, principally as an id to the
assessment of former production from mined-out deposits by tailings examination. Washing of gravel
piles by licensed miners has been observed, and their recoveries noted; the tailings have then been re-
treated by the Geological Survey team with its own equipment, and recoveries compared for various
diamond sizes. Altogether, about four hundred cubic yards of the better-grade gravel have been treated
in this way over the whole period, and the results have been fairly consistent throughout. The
conclusions are summarised in Table 1.



Size range of diamonds in carats Average losses during treatment of gravel by
licensed miners




The median diamond size was 1.0 carat, and the numerous stones of 0.2 carat and below
constituted only 8% of the total caratage. The heavy losses of very small stones are therefore not
particularly calamitous, especially when one remembers the low prices fetched by these stones.
Average losses over the whole size range were 14%

– 12 –

Over 2
1 to 2

0.5 to 1

0.2 to 0.5

0.1 to 0.2

0.05 to 0.1
Less than 0.05

Illicit miners employ the same extraction and treatment techniques as the licensed miners, but
with efficiency even further impaired by haste and night working. During the years 1950 to 1955, all
mining by native miners was illicit, but the introduction of the Alluvial Diamond Mining Scheme in
1956 brought most of it under control and it is now only prevalent in certain places. These are:-

1) In parts of the Yengema Mining Lease of Sierra Leone Selection Trust, where there are
well-established communities whose economics are based on illicit mining and dealing

2) In inaccessible areas, when a new discovery is made in a locality where mining has not
formerly taken place. The immediate result is a diamond rush and an outbreak of illicit
mining, but this is normally converted to licensed mining without undue difficulty on the
arrival of officers of the Mines Division.


The subject of mining costs and the payability of alluvial diamond ground is somewhat complex,
because of the wide variety of conditions, both natural and financial, which apply. A mining company
such as Sierra Leone Selection Trust, operating outside the alluvial Mining Scheme, is subject to
various taxes and is expected to provide houses, schools, hospitals and other amenities. Consequently,
operating costs are high. They tend to rise from year to year, but in 1962 for example, S.L.S.T.’s total
operating costs were Le.6.60 per cubic yard of gravel mined, in addition to capital expenditure, which
averages about Le.800,000 per annum.

Native Companies, formed under the provisions of the Alluvial Diamond Mining Scheme,
operate under more favourable conditions since they pay no taxes and incur only limited obligations to
welfare expenditure. Total operating costs of a small Native Mining Company (i.e. one treating about

50 cubic yards of gravel per day) are estimated to be about Le.5.00 per cubic yard for most alluvial
deposits, but can vary considerably according to the conditions. For example, the costs of mining an
extensive high terrace deposit should not exceed Le.3.50 per cubic yard, but the total costs of
constructing a large coffer dam on the Sewa River, and extracting and treating the gravel, using manual
methods, are about Le.10.00 per cubic yard of gravel washed.

Licensed miners have the lowest costs of all, since these merely consist of the fee for the licence
and the cost of simple equipment: picks, shovels, headpans, and sometimes a pump. Having paid for
these items, the licence-holder then has no further major expenses, irrespective of the number of
tributors he employs. He normally takes one-third of the sale proceeds of all diamonds recovered, and
his income is more dependant on the number of tributors he can attract than on the values present.
Having undertaken his initial expenditure, which is irrecoverable, he will wish the plot to be worked for
some return however regligible the values turn out to be, because no further outlay is entailed.
Whether the gravel is mined depends not on the licence-holder, therefore, but on whether the values
present will yield an acceptable income to tributors. Under optimum conditions, the output per
man/month is about 7 cubic yards of gravel extracted and treated. Normally, the minimum acceptable
income in the diamond fields is Le.15.0 per month, and in order to receive this, each man must produce

1.4 carats of diamond which will be sold for Le.22.50, the licence-holder taking one-third. The
minimum payable grade therefore, by this computation, is 0.2 carats per cubic yard and this
corresponds well lwith actual practice. Gravel with values lower than this is rarely mined, except
where large diamonds are expected; gravel with values of 0.3 carats or more usually attracts the full
quota of twenty tributors per licence. The values quoted apply, of course, to the gravel actually mined;
the miners can and do mine 0.2 gravel selectively from deposits with overall values as low as 0.08
carats per cubic yard.

– 13 –
The costs of suction dredging in a river channel can be reliably estimated from the data supplied
by forristal (1965). For a small suction dredge of the type employed at Boajibu, which treated both
sand and gravel, and had a throughput under optimum conditions of 25 cubic yards per hour, actual
running costs would be about Le.2.40 per cubic yard. In order to cover running costs, return of capital
and interest on capital, a minimum reserve of 500,000 cubic yards with an overall grade of not less than

0.2 carats per cubic yard would be required. Sites where a large bucket dredge could be employed,
stripping overburden before mining and treating gravel, are limited to Blocks 16 and 17 on the Lower
Sewa. There are as yet no data on which an estimate of costs for this type of operation in Sierra Leone
could be based.

The author’s estimates of minimum grades for profitable mining in Sierra Leone under the
Alluvial Diamond Mining Scheme are summarised in Table 2:-


Swamp or
Shallow flat

Deep flat

Deep low



Licensed Miners:
Minimum values in carats per
cubic yard of pay gravel

0.2 unsuitable unsuitable 0.2 0.4

Native Company:
Minimum values in carats per
cubic yard of pay gravel

0.3 0.4 0.35 0.25 0.5
dam or

Suction dredge: minimum
values in carats per cubic yard
averaged over all channel

unsuitable unsuitable unsuitable




In their prospecting, the alluvial diamond miners use the same techniques as they employ for
mining, the difference being simply one of pit spacing. The prospectors traverse the countryside
placing a few random pits in most of the swamps or stream flats they meet and extracting some bottom
gravel from each. A few headpans of gravel from each pit are tested by washing in the customary
perforated pan and jigging in a small hand jig. If corundum is absent or uncommon, the remainder of
the gravel is discarded and the swamp pronounced barren, unless of course, a diamond has been found,
which would be most fortuitous in such a small sample. If reasonable quantities of corundum are
found, then all the extracted gravel is washed. The discovery of abundant corundum or of a diamond
leads to numerous additional pits being scattered over the deposit; wherever more diamonds are found,
pits enlarge and proliferate and mining has begun. At some stage of the procedure described above, a
Mining Licence will normally have been taken out, sometimes before any diamonds are found.

– 14 –
The shortcomings of these methods are clear. They are:-

(a) Only when corundum happens to be present is an adequate sample taken. When corundum is
absent the quantity of gravel washed is completely inadequate to establish the absence of

(b) The use of corundum as an indicator is only justified where it is water-worn and hence indicates
a concentration point. Angular corundum signified nothing, but the miners do not make the

(c) Inefficient treatment methods result in the loss of 15% of the heavy mineral concentrates, and
these will sometimes include the only diamond in the sample.

(d) The unplanned and unsystematic nature of the work mean that although complete cover of an
area may sometimes be achieved, the cost in man hours is tremendous by any normal standards.
This particular objection is, of course, of no consequence to unemployed miners who are
prospecting on their own behalf and do not count their labour as cost.

In spite of these many shortcomings, however, the method has been outstandingly successful in
the past in south-east Sierra Leone because corundum is of widespread occurrence and the majority of
the know deposits are at least partly alluvial in origin, consequently most diamond concentrations are
associated with corundum concentrations. This, however does not apply to alluvial diamond
concentrations, neither does it apply to diamond concentrations of any type in those parts of Sierra
Leone where corundum is deficient. Another reason for the past success of the method is that, in the
vicinity of any productive deposit, prospecting activity becomes so intensive that any payable deposit
must be found eventually however inefficient the method. This leads to steady expansion of the mining
field until the limits of diamond occurrence are reached. It is only within and on the margins of the
known fields that such saturation prospecting occurs; it does not and cannot happen elsewhere because
of the absence of the necessary large numbers of itinerant alluvial miners.


The prospecting procedure summarised here are those now followed by the Geological Survey,
and are regarded as being adequate to find even minor alluvial diamond fields under Sierra Leone
conditions. They are based on the methods formerly employed by Sierra Leone Selection Trust during
their prospecting operations, but they have been somewhat modified in the light of recent experience
with small isolated alluvial fields.

(a) Treatment of gravel

All gravel extracted during prospecting operations should be treated as follows to recover the
diamonds present:-

Process 1 – Sizing of gravel and removal of fines. This is accomplished by passing the
gravel through a sequence of sieves, manually operated under water in a stream-channel
or flooded trench. Alternatively, and more efficiently this operation can be carried out in
a foot-operated rocker unit, which consists of a vertical series of nesting screens in
wooden frames, into which water is discharged by a pump. Properly operated rockers
have a higher throughput than manual sieves, but are somewhat less portable.

Process 2 – Jigging of sized gravel. This is carried out in a “gravitator”, a circular frame
of 15 inches diameter on which is stretched a 1mm phosphor bronze screen. The
gravitator is simultaneously jigged and rotated by hand at the water surface. Heavy
minerals descent and are centralised on the screen.

– 15 –
Process 3 – Diamond picking. After being allowed to drain, the gravitator is inverted on a
level surface and then lifted, leaving a flat cake of gravel with the heavy minerals exposed
at the centre of the upper surface. Diamonds are picked off by the prospector only. When
heavy minerals are abundant in the central “eye” they should be scooped off and collected
for re-jigging later.

Being a manually-operated system, the sieves require very close supervision by the
prospector to avoid theft and losses through careless operation. In spite of precautions 100%
recovery can rarely be achieved. Furthermore, only about 5 cubic yards of gravel can be treated
in a day by ten men, and consequently where much gravel is to be treated at one site the use of
rockers and Joplin Jigs is often to be preferred. However in basic prospecting the primary
consideration is frequently the portability of the equipment and the rapidity with which a
washing site can be brought into operation. From these points of view the sieve-gravitator
equipment cannot be bettered and the consequences of infrequent diamond loss can be
minimised by taking samples of appropriate size. The perfect system does not exist.

(b) Location of a diamond field

In this, the first stage of reconnaissance prospecting, samples are taken at approximately
five mile intervals throughout the drainage network, along all streams and rivers of the 5th order
and above. Additional samples are sited at suitable points on 4th order streams in intervening
areas to achieve a final overall sample density of at least one sample per 9 square miles. At each
sample site, the prospector washes at least 20 cubic yards of enriched gravel identifiable by the
fact that its heavy mineral content exceeds the average for the gravels of that particular reach of
the stream. Where practicable, the gravel is collected from suitable parts of the stream channel,
but in the case of some of the smaller streams pitting of the flood-plain may sometimes be

(c) Location of deposits, and delimitation of the field

If two or three positive samples have been produced during the first stage, a region of
alluvial diamond occurrence has been indicated. If the positive samples occur in succession
along a drainage route, the area of greatest interest is that drained by the tributaries entering in
the vicinity of the furthest upstream positive sample. In the second stage of the exploration, the
region indicated by the results of the first is further examined by means of samples taken at
intervals of 2.000 feet throughout the draining pattern. Each sample consists of at least 10 cubic
yards of basal gravel taken from a minimum of five pits sited in the swamp or stream flat. The
precise arrangement of the pits is unimportant provided they are at least 100 feet apart and
provided no attempt is made to group them in an area of abnormally enriched travel. A roughly
representative sample of the flat gravel is required. Channel or rock-bar gravel is unacceptable
at this stage. These methods will locate all major deposits in the field and will give an
approximate indication of the values present.

(d) Evaluation of deposits

All stream flats which during the second stage of the exploration have been shown to be
diamondiferous are further developed by systematic pitting in accordance with normal alluvial
valuation principles. Where the flats are found to be payable pit lines are extended on to the
terraces. At this stage every effort must be made to eliminate diamond losses, so sieve
equipment, if still in use, must now be superseded by rockers.

Part I. Section F – THE DIAMONDS


A detailed description of diamonds from Koidu has been given by Grantham and Allen
(1960) and the characteristic features they record occur throughout the diamond fields.

– 16 –
Although there are marked differences between the production of separate fields, these are usually
differences only o proportion of the various types. In all parcels, the most important distinction to be
made is between the clear and the coated stones.

Most of the clear stones are colourless and of good quality, but a few yellowish, pale green and
pale brown stones occur in some deposits. About 20% of the clear stones contain small black
inclusions, often surrounded by radiating cracks and in a few cases the inclusions are so abundant as to
render the stone valueless as a gem. The dominant crystal form is the octahedron, usually slightly
flattened and sometimes twinned or distorted. Doceahedra are rare and cubes do not occur among the
clear stones. The octahedra may exhibit various growth features such as plates, layers, shields or
crinkling but a good proportion, especially from the Yenegma and Koidu areas are “glassy”, with plain
faces. The corners and edges are sometimes chipped or worn even among stones taken from

The coated stones vary in external colour from pale greenish-grey to dark green and they
normally consist of a clear inner stone covered by a layer of translucent green diamond. This coating
may be as much as 3mm thick, although the average is about ½mm in crystal form. Also, the coated
diamonds differ from the clear being generally regular symmetrical octahedral, dodecahedra or cubes,
with faces unmodified by plates or shields and exhibiting rough or matt surfaces. In many cases the
crystal form assumed by the whole coated stone is different from that of the inner clear diamond.

The relative proportions of clear and coated stones vary from one deposit to another. In addition
irregular shaped grey-black bort occurs in small quantities in most deposits. The overall ratios in the
production from the principal fields are estimated to be as shown in Table 3.


Percentage by weight

Clear Coated Bort

Yengema/Field 50 45 5

Tongo Field 93 3 4

Upper Sewa 58 38 4

Middle Sewa 66 30 4

Lower Sewa 77 20 3

Matemu 90 2 8

Kenja/Moa 97 1 2

The range of sizes over which the Sierra Leone diamonds occur is a large one, and in any
standard month’s production the weights of individual stones will vary from 0.03 carat to 50 carats
from the Yengema area. At the other end of the scale, a few perfect octahedral which are estimated to
weigh as little as 0.002 carat each have been detected in the fine fraction of pan concentrates from
weathered kimberlites.

The question of average diamond weight which interests many people, is not an easy one to
approach. In the first place essential data are lacking because diamonds are never counted and
classified by weight in the sorting rooms or the buying offices. In the second place there are different
methods of calculating the average. Average weight is usually assumed to be the Arithmetic Mean of
individual weights, but this is in fact a most misleading figure to employ since the very small diamonds
have a disproportionate effect on it.

– 17 –
In most large parcels, for example, diamonds of less than 0.3 carat each though forming less than a
tenth of the total caratage, may nevertheless represent over a quarter by number. A far more useful and
representative figure is the Median weight, which appears in all references to average sizes in this

The following approximations have been calculated from the data collected during the present


Mean Weight Median Weight

S.L.S.T. Production (Yengema plus Tongo)

0.63 ct.

1.1 ct.

Alluvial Diamond Mining Scheme production

0.66 ct.

0.9 ct.

Bulked Sierra Leone Alluvial Diamond production 0.65 ct.

1.0 ct.

Koidu kimberlites

0.13 ct.

no data

Tongo kimberlites

0.14 ct.

no data

Panguma kimberlites

0.12 ct.

0.4 ct.

The licensed miners have their own terminology for classifying the diamonds they recover. Coated
stones are referred to as “orange colour” if pale or medium green, and “coffee colour when dark green.
Grey-black bort is termed “burnt” diamond. Stone sizes are visually estimated in accordance with an
arbitrary scale of numbers, which can be correlated with the actual weight of the stone in carats as

Jack More than 25 carats

Big 4

2 to 2.8 carats

Number 1 14 to 25 carats

Full 4

1.4 to 2 carats

Full 2 10 to 14 carats

Number 4 0.8 to 1.4 carats

Number 2 7 to 10 carats

Second 4 0.4 to 0.8 carats

Full 3 5.5 to 7 carats

Number 5 0.1 to 0.4 carats

Number 3 4 to 5.5 carats

Number 6 Less than 0.1 carat

Second 3 2.8 to 4 carats


There are three separate price levels for Sierra Leone diamonds; the prices received by the
licensed miners, the prices received by the dealers when the diamonds are sold to the Government
Diamond Office, and the final export value after all taxes have been paid.

In the 15 years since alluvial diamond digging began in Sierra Leone, world diamond prices
have rised to a limited degree. The prices received by the alluvial miners however, have increased
about tenfold over this period and this has corresponded with a steady reduction in the profit margins of
the dealers.

In 1954 the miners were losing a very high proportion of the smaller stones and were discarding
most of the coated stones as unsaleable. Their production was consequently of a high average quality,
and they were receiving an average price of about Le.6.00 per carat. It is estimated that this production
was finally sold on the outside market for about Le.25.00 per carat, and that its true export value at that
time was about Le.32.00 per carat.

By 1960, smuggling had been drastically reduced and the miners had a better idea of the value of
their production. The miners now recovered the coated stones and many of the smaller stones, but for
this lower quality production they were able to obtain an average price of Le.14.00 per carat.

– 18 –
Dealers received an average price of Le. 19.0 from the G.D.O. but some of the better-quality gems
were still being smuggled, so the overall price received by the dealers must have been higher, probably
about Le.23.00 per carat. The final export value of the G.D.O. purchases at this time was Le.20.00 per
carat, but the value was of course, depressed by the absence of some of the better gems.

In 1965, the miners have been receiving an average price of Le.20.00 per carat for their
production whilst the G.D.O. has been buying at Le.28.00 from the dealers. The final export value was
approximately Le.29.50 per carat.

The export value of the Sierra Leone Selection Trust production is about Le.24.00 per carat.
Two factors appear to be responsible for the difference between this figure and the value of the
A.D.M.S. production:-

1) The proportion of coated stones is higher in Yengema Field than in any part of the alluvial
miners’ areas.

2) The proportion of small stones is higher in the S.L.S.T. production because of losses by the
alluvial miners during gravel treatment.

In this Bulletin it has been assumed for the purpose of valuing resources, that the following
export values will apply:-

Diamonds recovered by dredging or mechanised mining (chiefly from Sewa deposits)
Le.24.00 per carat

Diamonds recovered by licensed miners
Le.28.00 per carat

It has further been assumed that the production of mining companies, however constituted, will be sold
direct to the G.D.O. and not to dealers, and the average price realised should therefore be about
Le.22.00 per carat. This figure has been used when assessing the payability of deep ground.


Sierra Leone Selection Trust production in 1965 was 652,410 carats, and the Company’s average
annual production for the last ten years has also been about 650,000 carats.

Production under the Alluvial diamond Mining Scheme in 1965 was 809,544 carats. This figure
is slightly above 1963 and 1964 levels because recent increases in Central Selling Organisation price
levels have resulted in the virtual elimination of smuggling.


Revenue from diamond mining is an important factor in the economy of Sierra Leone and in
recent years has averaged Le.4.8 million per annum. Part of this revenue is attributable to the Alluvial
Diamond Mining Scheme and part to Sierra Leone Selection Trust.

Revenue from the A.D.M.S. has consisted of the 5% export duty* on diamond sales, together
with various licence and permit fees. During the last few years the average annual revenue from these
sources has been about Le.1.3 million, which represents Le.1.75 per carat produced, or about 7% of the
export value of the diamonds.

Revenue from S.L.S.T. consists of the Diamond Industry Profits Tax, Income Tax, the Service
Fee on diamond sales, mining lease rents, and annual payments to Kono District Council and Lower
Bambara Chiefdom.

* The export duty has now been increased (1966) to 7½%. Consequently total revenue from the
A.D.M.S. should now increase to about 10% of the export value of the diamonds.

– 19 –
These headings have averaged Le.3.5 million per annum, which represents Le.5.10 per carat
produced, or approximately 27% of the export value of the diamonds.

Detailed revenue figures, on which the above generalisations are based, are published in the
Annual Reports of the Mines Division.

– 20 –



South-East Sierra Leone is part of the West African craton, and almost all the diamond fields are
therefore underlain by Pre-Cambrian basement formations, predominantly granitic in type. Detailed
mapping of certain sections has been undertaken by the Geological Survey in recent years and the
results have been recorded in Bulletins 1, 2, 4 and 6. Interested readers are therefore referred to these
and to Pollett (1952). A brief summary only will be given here.

The formations which occur in south-east Sierra Leone are listed in Table 4.



Formation Description Distribution within

S.E. Sierra Leone

Clay, silt, sand,
Clay, silt, sand,

River valleys

Quanternary (recent alluvium

River Terrace deposits)


Bullom series Unconsolidated
sediments chiefly silts

Coastal belt only. No
diamonds yet found on
or in Bullom Series

Kimberlites Dykes of prophyritic
kimberlite and pipes of
kimberlite breccia

Yengema/Koidu area
Tongo/Panguma area


Dolerite dykes Normal ophitic
Some with olivine

Found throughout
(except in areas
covered by Bullom)


Early Palaeozoic
Late Pre-Cambrian

Rokell River Series Folded, unfossiliferous
ediments & volcanics

Belt crossed by Sewa
River south of Bumpe


Meta-dolerites Amphibolite dykes and

Widespread. Exposed
in channel of Lower

Late Pre-Cambrian Kasila series Hornblende + pyroxene
gneisses, granulites,

Belt running from
Sumbuya to area S.E.
of Zimmi

Massive unfoliated
intrusive granites,
usually hornblendic

Found cutting
synkinematic granites
in all mapped areas

Early Pre-Cambrian Late-kinematic


Predominant rock-type
Forms bedrock of most
diamond deposits
Early Pre-Cambrian Kambui Schists Pyroxene & hornblende

Kambui Hills.
Nimini Hills. Common
as strips in
granodioritic geniss.

Grano-dioritic gneiss
(Enderbite in Gola

Early Pre-Cambrian Syn-kinematic


Granulite belts
common to east of
Kambui Hills.

Early Pre-Cambrian Mano-Moa


– 21 –
Wilson (1965) also records numerous other rock types in the Gola Forest area: plagioclasites,
lamprophyries, diabase, vogesites,nepheline syenite, olivinite, chromitite and various minor ultrabasic

Throughout the diamond fields, the rock most frequently encountered in outcrop and boulders is
granodioritic gneiss, and this gives a misleading impression of uniformity. It has been observed that
bedrock exposed in both swamp and hillside trenches very frequently includes a high proportion of
other formations, such as amphibolites, schists and quartzites, which are present as strips and lenses
within the granite, but of which there is no indication in the form of outcrop or float. The geology of
the granite areas is far less uniform than would appear from the map colouration.


Schist belts, granulite belts and the foliation of the granodiorite gneiss are normally parallel in
any given region, all following a regional trend. Proceeding from north to south, these regional trends
show a progressive swing. In the north of the diamond fields, in Kono, the mean regional strike is 0°
true, and predominantly north-south strikes continue to be displayed as far south as Panguma, where
the progressive swing begins. In the Kenema-Bo region the regional strike has become 25° true, and
further to the south still, around Jimi, Potoru and Zimmi, it is about 40° true. Throughout the diamond
fields, cleavage and foliation are normally vertical or near-vertical.

The schist and granulite belts are considered to represent the roots of complex fold structures (cf.
Bulletins 1, 2 and 4) but, as the detailed succession within these formations is obscure and dips are
usually vertrical, actual folds cannot be demonstrated, except on very minor scale. The formations of
the considerably younger Rokell River Series, which are unmetamorphosed and structurally less
complex, and tightly folded along an axis running appro9ximately 160° true.

The granites are cut by numerous fracture systems which exercise close control over the
drainage pattern. Mylonites have been noted in many of the fracture zones, and shearing of the granite
is common, together with the development of pegmatities and course boitite clots. Faults and shear-
zones are always near-vertical, but only in a few cases can the amount or direction of displacement be
deduced. It is probable that displacement has generally been vertical and that most major fractures are
high-angle thrusts. A number of different fault directions may be observed, and no single direction
predominates, but, throughout the diamond fields, the most important and outstanding structures are the
strike faults.

Nearly all known kimberlite dykes have strikes of about 65° to 70° true and are steeply-dipping
or vertical. Most dolerite dykes have strikes which are somewhere in the range between 110° to 130°
true, although minor dykes with strikes in various other directions have been recorded near Koidu by
Grantham (1960), and have been noted near Bo in the course of the present survey. Although some
dolerites are vertical, many have dips as low as 65°.



The great majority of the diamond deposits with which we are concerned in this Bulletin are
superficial, and an appreciation of the morphology of the land surface is therefore essential to an
understanding of the deposits. A detailed discussion of the geomorphological processes would be out
of place here, but it may be appropriate briefly to note the principal factors controlling the operation of
these processes in the diamond fields.

– 22 –
Probably the most important factor is the climate, which is hot and wet with pronounced wet and
dry seasons. The high temperature and rainfall result in a dense vegetation cover, which in turn inhibits
the mechanical erosion of slopes and add organic acids to the influent seepage. The contracted wet and
dry seasons cause a considerable annual fluctuation in the water table, hence a deep zone is exposed to
the decomposition and solution of minerals by migrating ground water. Solution transport is the
dominant erosional process, and the weathered zone is therefore leached of silica and residually
enriched in the less soluble oxides, principally those of iron and aluminium, which form laterite
nodules and sometimes duricrust.

The prevalence of chemical weathering and solution transport results in a deficiency of whole-
rock fragments in gravel accumulations. Residual gravels may be thick, but consist largely of laterite
nodules. Alluvial gravels are thin, and are composed almost entirely of quartzite or vein-quartz
fragments, all derived from veins and lenses which form less than 1% of the total bedrock volume.
Shortable of abrasive load is an important factor in river profile development.

Another important factor is the nature of the commoner rock formations, their mineralogy,
texture, disposition and structure. By far the commonest rock is granite, which, when sound, is very
resistant to mechanical attack, but which is very susceptible to chemical attack wherever the
groundwater can obtain access. The granites are massive, and cannot be said to have any attitude, but
they are cut by numerous faults, shear-zones, and other lineaments whose exact nature is unknown, all
of which are reflected in the topography. Most structures are vertical or near-vertical, but horizontal
joints have been observed in the underground openings at Koidu. Schist bodies and dolerite dykes are
also generally sub-vertical.

Earth movements have also played an important part in the morphology of the land surface.
There is clear evidence of tilting and uplift, and this will be discussed later, but there is no evidence to
show whether any of the major fractures were associated with the uplift. Certainly there is no reason to
believe that fault scarps played any part in the evolution of the present topography.


As the greater part of the diamond fields lie on granitic rocks, the drainage throughout tends to
have a trellised pattern and to be closely related to structure. Rivers, streams, swamps, gullies; all
usually follow lines of weakness in the granite. Drainage density averages about four miles of
permanent drainage channel per square mile. This dense network is a result of the heavy rainfall in
conjunction with the groundwater conditions which obrain in granite terrain, where most groundwater
is stored in numerous minor shallow perched bodies in the weathered zone.


In a publication such as this, which is concerned with an entire alluvial mineral field, there are
inevitably references to a large number of rivers and streams. In most cases, various characteristics are
described, such as the gradient and the incidence of terraces, but it is generally also desirable to give
some idea of the size of the stream and its position in the drainage network. Width and mean annual
discharge are both unsatisfactory criteria for this purpose; width because there is such a great variation
within any given river segment, and discharge because in only a few cases has a stream been visited
sufficiently often to enable a useful estimate to be made.

It has therefore been decided here to adopt the system of Stream Order devised by Horton
(1945), as modified by Strahler (1952). This has the advantage not only of defining unequivocally a
stream’s status in the drainage pattern, but also of requiring only allotment of a simple number which
can be readily determined from the map or air photo and which involves no subjective element of
judgement or estimation of quantities. Furthermore, within the limits of a region such as south-east
Sierra Leone, steams of the same order tend to have similar average widths and discharges.

– 23 –
In the Strahler systems of Stream Order, a first-order channel is one which has no tributaries i.e.
a “fingertip” channel. A second-order channel is formed by the confluence of two first-order channels,
and has as tributaries only first-order channels. When the second-order channel lis joined by another, it
becomes a third-order channel. A third-order channel joined by another becomes fourth-order, but if
joined by a channel of lower order it remains third-order. And so on, with the stream order increasing
as one proceeds down the profile.

The drainage network above a given point can be defined by reference to the stream order at that
point. For example, in the south-east quadrant of Sheet 58, the Bandafaiyi (5th order) joins the middle
Gbobora (5th order) to form the lower Gbobora (6th order). The Bandafaiyi and tributaries are thus a
fifth-order network but the complete Gbobora system, which includes the Bandafaiyi, is a sixth-order

In some territories the order given to a stream segment must vary with the scale of the map used
to calculate it, since a large-scale map will show more gullies and other ephemeral channels than a
small-scale map, and all must be included in the network, thus increasing the orders throughout. This
difficulty is avoided in south-east Sierra Leone by regarding the swamps as the outermost members of
the network. Gullies feeding swamps are excluded unless there is perennial surface flow, and this is an
extremely rare condition, as springs are generally found only at the heads or margins of swamps.
Correct identification of first-order channels may sometimes be difficult on the 1:50,000 maps, but it
can generally be done with ease on the air photos, except in heavily forested regions.

Within south-east Sierra Leone where the average annual rainfall in about 90 inches, the general
characteristics which apply to nearly all stream channels are those shown in table 5.



Average distance
from beginning
of segment to
principal source

(cubic feet per


Proportion of
smooth well-
rounded quartz
pebbles in
+6mm. fraction
of normal
channel gravel





250’ or 3’






Usually swamp
Rarely stream

Occasionally swamp
Usually stream

Stream e.g. Gbatiye
at Nongoba

Steam eg Bakwe at
Sembehun 17

Small river eg
Lower Moinde

River eg
Matemu at Gbado

River eg Woa at

Large river eg Bafi
At Yomadu Mal
Principal river
e.g. Sewa, Moa









Rare, large only


5 miles




10 miles



1% to 4%


18 miles



2% to 10%


28 miles



6% to 15%


45 miles

100 miles

3 (mostly

12 (partly






12% to 25%

18% to 60%

The proportion of perfectly-rounded pebbles tends to be higher in flood-plain gravels, terrace
gravels and potholes. However, the occurrence of rounded pebbles in excess of 10-% in any gravel of
fifth or sixth order streams, or of any rounded pebbles at all in streams of lower order, indicates that the
stream in question is reworking gravels deposited originally by an ancient drainage of higher order.


Rarely can the classical concave profile be recognised in south-east Sierra Leone. Frequently
the headwaters of a major river have a gentler gradient than the middle and lower reaches. In detail, all
rivers evolve a stopped profile consisting of long sections of shallow gradient (0.5 to 1.0 feet per mile)
separated by short sections of steeper gradient, (10 to 50 feet per mile).

The sections of steeper gradient are of two types:

1) True nick-points, which retreat up the profile. At these points, the river descents abruptly from
one regional planation surface to another.

2) Permanent breaks-of-slope, tied to the more resistant zones in the granite.

At the true nick-points, a substantial drop in elevation occurs within a very short distance.
Examples from the major rivers are:-

The Sewa near Pundura, where the river drops 300 feet in 2 miles (including Big Kongo falls).

The Moa near Kailahun, where the river drops 400 feet in 16 miles

The Tye at Palima, where the river drops 300 feet in 3½ miles.

At all these points, the rivers are dropping from the Thousand-foot Surface to the Coastal Plain surface,
and the nick-points appear to be retreating rapidly up the profile.

The permanent breaks-up-slope generally take the form of shallow rapids or rock bars, and their
existence is due to zones of more resistant rock, generally rock with fewer faults and joints. In the
basement granite, such zones tend to be disposed vertically rather than horizontally, consequently the
breaks-of-slope tend to remain at a fixed position in the profile.

The sections of shallow gradient can be regarded as graded, in that a condition of virtual stability
has been reached. It is a matter of observation that meanders and floodplains have often formed along
the smaller rivers but rarely occur along the larger. The reason for this is that the graded sections tend
to be of similar length for all rivers, since the distance between resistant bedrock zones is independent
of the size of the river which flows over them. In the case of the major rivers, the ratio of this graded
length to channel width is generally small, less than 20:1, and channel swinging is thereby inhibited.

Generalised profile gradients of the major rivers of the diamond fields, travelling upstream from
the coast are:-

Sewa/Bafi system – Below Sumbuya –

0.5 ft/mile(for 16 miles)

Sumbuya to Gerihun

7.4 ft/mile (for 34 miles)

Gerihun to Punduru nick-point –

4.1 ft/mile (for 56 miles)

Punduru nick-point

150 ft/mile (for 2 miles)

Punduru to Bafi confluence –

20 ft/mile (for 12 miles)

Lower Bafi

4 ft/mile (for 30 miles)

– 25 –
Moa Coast to Bogboabu

Middle Moa

Upper Moa (Kailahun section) –

Upper Moa (Koindu section) –

It would appear that, after nay disturbance such as tilting of the land surface or rapid change of
base level, a river flowing on granitic rocks soon regains a condition of sub maturity, with a mean
gradient of about 4 feet per mile over long distances. Further reduction of the mean gradient would
require elimination of the permanent breaks-of-slope, and a very long period of base-level stability
would be necessary to bring this about.


Inspection of the topography in the vicinity of the major rivers shows that in most cases there are
a number of terrace levels, and these are particularly apparent in the valley of the middle Sewa. The
succession there, from youngest to oldest, is as follows:-

1) Recent Flat or Flood Plain

2) Low Terrace. This is a few feet above the flat, has deep overburden and is generally un-

3) First High Terrace. This is generally found flanking the low terrace or the flat, and may be from

10 feet to 40 feet above the low terrace level. It is usually lightly dissected, and much of the
terrace alluvium has been removed by erosion, but recognisable alluvial gravels or gravel
residues remain. Where the terrace is cut by swamps, their gravels contain abundant re-
concentrated terrace material, including diamonds.

4) Second High Terrace. This is a zone of low flat-topped hills and ridges, separated by swamps,
which flanks the river on both sides. The concordant flat hilltops clearly indicate a former
terrace level, which is from 50 to 80 feet above the present flood-plain level. Alluvial residues
can rarely be detected in the soil, but the former existence of terrace deposits is confirmed by the
presence of water worn heavy minerals in all the swamps, and by an occasional solitary rounded
quartz pebble. A few diamonds are found in most of the swamps which cut this terrace, and in
some cases these constitute workable deposits. diamonds are rarely found in the lateritic gravels
of the hill slopes.

5) Third High Terrace. This is a wide zone of subdued topography with concordant hillcrest
elevations only 20 or 30 feet above the Second High Terrace level. Along the middle Sewa, the
third High Terrace extends to the limits of the valley floor, and in the Blama area it merges with
the main expanse of the Coastal Plain Surface. Generally speaking, although the swamp gravels
often contain traces of water worn heavy minerals, they contain no diamonds. In the rate cases
where they do, the diamonds are either believed to be alluvial or are associated with a tributary

Full discussions of the alluvial deposits found on the flood-plain, low terrace and first high
terrace levels are given in section C (ii). All these levels are present along the Sewa only as narrow
intermittent strips and they are here regarded as a single group, formed during the incision of the
Coastal Plain Surface, by the major rivers, in the period of lowered sea-level and fluctuating climatic
conditions associated with the Quaternary Ice Ages.

The second high terrace, on the other hand, represents a former continuous flood plain, about
two miles wide, formed on the coastal Plain Surface during a long period of stability before incision
began. Its age can therefore be guessed as late as Pliocene. There are few alluvial traces within the
third high terrace zone, and it is by no means certain that it does represent a fluvially-planed surface.

– 26 –

6.8 ft/mile (for 50 miles)

3.8 ft/mile (for 70 miles)

25 ft/mile (for 16 miles)

12.8 ft/mile (for 18 miles)

Swamps constitute 80% of the drainage network in the diamond fields. Almost all first and
second order drainage channels are swamps, and some of the third also. Alluvial diamond mining sites
are colloquially referred to as swamps and most of them are.

The swamps tend to be long and narrow, and are generally flanked by convex slopes. Actual
widths vary considerably, but the average is about 250 feet, while lengths of individual segments are
usually between 1,000 feet and 3,000 feet. It is rare for any part of a swamp network to be more than
one mile from the perennial stream which it feeds. Within a swamp the surface is flat and moist, and a
typical soil profile would be :-

0 to 2’ : Black mud with a high content of organic debris

2’ to 5’ : Grey clay, silt or sand

5’ to 5’9” : Bleached angular quartz gravel, sometimes in a clay matrix

From 5’9” down: Decomposed bedrock. Typically a stiff white clay containing
Kaolinite and sericite and irregular quartz granules, after granite.

Granitic bedrock is usually thoroughly kaolinized to a depth of about 10 feet below bedrock
surface, and the upper half of this kaolinized zone is always bleached whilst the lower half is often
orange stained. The cause of the bleaching of the gravel and bedrock is not immediately apparent, but
the phenomenon is not confined to swamps: it occurs also below flood-plain deposits. Where the
bedrock is schist, or much-fractured granite, complete decomposition to depths of 50 feet and more has
been noted and the resulting clays are usually discoloured and stained rather than bleached. One of the
commonest schist types to occur as strips and lenses within the granite is a fine-grained amphibolite of
plagioclase, hornblende and biotite, and under the swamps this rock becomes a sticky plastic sericite-
rich clay. This may be a uniform grey, blue, or blue-green at the top but becomes increasingly brown
and variegated at depth. Gravel lying on schist bedrock is usually grey.

On the surface, a small stream may flow along the swamp, but it is never deeply incised and
rarely makes contact with the gravel bed. Under no circumstances does the stream transport gravel or
erode bedrock. Often, in the dry season, the stream disappears entirely; conversely, many swamps
become flooded during the rains, because low gradients and dense vegetation permit only a very slow
rate of discharge. The water table is normally close to the ground surface, which consequently remains
moist, and although the water table may sometimes fall several feet during the dry season, it never falls
below the bedrock surface.

Many swamps, at their heads, enlarge to wide triangular or circular flats in which no stream or
stream course is ever visible and flow is entirely sub-surface. Such instances provide a clear
demonstration of the fact that swamp flats and floor-plain flats are fundamentally different features.

Swamps are formed by the corrosive migration of seepage points along outcropping faults or
joints where, under Sierra Leone conditions, the rocks are deeply rotted and offer little resistance to a
process of mechanical alluviation. The un-resistant nature of the fracture zone means that whilst the
seepage eats its way forwards, it has no tendency to migrate upwards, and the slight upward component
of its path is merely that necessary to ensure that the watercourse it leaves behind it has the minimum
gradient capable of conducting the seepage water away. A graded channel, a phenomenon usually
found in the lower parts of a stream profile, is thus formed at the source. Intermittent wet-season
seepage causes basal sapping of the blanking slopes, widening the watercourse and producing a waster-
storage basin in which water movement is very slow indeed. The accumulation of gravel, fine alluvium
and organic debris in this basin result in the formation of a swampy flat.

– 27 –
The coarser residues, from the rock masses disintegrated by the migrating seepages, cannot be
transported by the slow-moving swamp flow and consequently suffer virtually no horizontal
displacement within the swamp system. They form the basal gravels of the swamp which are thus seen
to be residual and not alluvial. On the other hand, the over-lying silts, clays and muds can be
legitimately described as alluvial since they have been sorted and deposited either by a small
meandering swamp stream or by wet-season flood waters.

In a majority of headwater swamps, at the height of the dry season, seepage discharge becomes
minimal and takes place directly into the basal swamp gravel. Water flow along the swamp is then
entirely sub-surface and may remain so for several hundreds of feet.

Laterite in various forms is the commonest superficial formation in the diamond fields, and it’s
presence at any locality testifies to the past feebleness of the processes of mechanical erosion. The
term is often used loosely to denote material of lateritic affinities, but it is preferable to employ more
specific names. Lateritic clay is a red-brown iron-rich alluvial clay. Lateritic nodules are small
concretions of hydrated ferric oxides and alumina. Leteritic gravel consists of abundant lateritic
nodules in a lateritic clay matrix.

True laterite is a hard brown porous formation composed of hydrated ferric oxides and alumina,
with a little silica, and is essentially a residue resulting from the removal of the relatively soluble silica
from weasthered rock by groundwater. It forms a crust at the upper surface of a weathered rock mass
and thickens downwards by replacing the weathered rock. Thick sheets of laterite are often referred to
as “duricrust”, and are normally found only on extensive level land surfaces. Although no exact
relationship can be defined because of the varying underlying rock-types and possibly varying past
climatic conditions, there appears to be a rough correlation between laterite thickness and age of the


The principal topographical divisions of south-east Sierra Leone are the coastal plains and the
inland plateau. Both represent planation surfaces of major importance, but the southern scarp of the
plateau is split into a series of steps or minor surfaces. The southern margins of the plateau also
include extensive areas of rugged dissected topography where no planation surface can be recognised,
although the probable level of the destroyed surface can often be inferred from concordant flat
summits. On the plateau there exist small high remnants of earlier surfaces.

The surfaces noted during the present survey are shown on May III. Proceeding from the
highest to the lowest, they are:-

I. The Nimini Surface. Within the diamond fields, this surface is represented only by the
undulating duricrusted area which caps the Nimini Hills west of Jaiama Nimikoro. According to the
old 1:62, 500 Topographical Sheets, the average elevation of this surface is about 2,250 feet. This
solitary high remnant is, for reasons which will be discussed later, provisionally correlated with other,
more extensive, high surfaces outside the diamond fields. No diamonds have been found on it.

II. The Main Plateau Surface. This, with an average elevation of about 1,500 feet, is one of the two
principal surfaces of Sierra Leone, but in the diamond fields it exists as a recognisable surface only in
the extreme north. Between latitudes 8°15’N and 8°45’N it is now represented only by large areas of
dissected hill country with many flat summits and with summit elevations all between 1,400 feet and
1,600 feet. Isolated laterite blocks can occasionally be found on hill slopes in this dissected country
and are assumed to be fragments of extensive duricrust sheets on the recently-destroyed surface. Very
few diamonds have been found on the Main Plateau Surface.

III. The Koidu Surface. This surface has an elevation of 1,250 feet and is oddly situated. Itself of
exceptional flatness, it is almost surrounded by hilly country representing higher surfaces and thus
forms a basin-like feature in which great depth of bedrock weathering is general. The Koidu surface is
apparently only of local importance, having few equivalents elsewhere. A majority of the known
kimberlites crop out on it and the most important alluvial deposits lie on it.

– 28 –
IV. The Thousand-foot Surface. This has been so named because of its consistent elevation. It is of
considerable importance within the diamond fields as it can often be traced without break over long
distances, and it is generally apparent as the principal step or bevel between the Coastal Plain and the
dissected margins of the Plateau Surface. Inselbergs are sometimes numerous but, between them
duricrust up to 4 feet in thickness is often found capping the low interfluves. A small number of
diamond deposits occur on this surface.

V. The Tongo Surface. North of Kenema, the important diamond deposits of Tongo field lie on a
low plateau or shelf with a consistent average elevation of 750 feet. It is of limited extent, occupying
an embayment in the fringes of the Main Plateau, and is separated from the slightly lower Coastal Plain
surface by an escarpment which is clearly recognisable in the field, although somewhat convoluted.
The Tongo relief is low and undulating, similar to that which typifies the Coastal Plain Surface, but
duricrust is rarely seen.

VI. The Coastal Plain Surface. This surface occupies over half of Sierra Leone, and in terms of area
(but not of resources) the greater part of the diamond fields lie on it. The elevation varies with the
distance from the coast, being about 100 feet at the outer margins, 400 feet at Blama, 500 feet along the
Upper Sewa and 600 feet along the Upper Moa. There is a fairly consistent gradient of about 4 feet per
mile towards the coast. This is considered not to indicate tilt but rather to represent the natural gradient
of a fairly youthful surface as determined by the stable profile gradient of sub-mature drainage on
granite. Relief on this surface consists of numerous, low, convex or flat-topped hills separated by a
network of swamps. Isolated hills of inselberg type occur in many parts, but are only common north of
Bo. Duricrust is thin and sporadically distributed, but decomposed bedrock is generally mantled by a
thick zone of laterite gravel. The lack of topographical variety on this surface gives the impression that
much of it is of marine origin and such an interpretation is in fact probably correct, as there is now
general agreement that Late Tertiary sea-levels throughout the world were at least 400 feet higher than
at present.

VII. The Bullom Surface. This is the most recently initiated surface, and forms a coastal strip up to

25 miles in width, with elevations from 0 to 50 feet. It corresponds roughly in extent with the outcrop
of the unconsolidated Bullom Series, and is in part of depositional rather than erosional origin. No
diamonds have been found on any part of this surface. Relief is flat or very mildly undulating.


The existence of the step-like series of planation surfaces will not be disputed by anyone who
has worked extensively in the field in south-east Sierra Leone, but there will doubtless be some
disagreement over their interpretation, and it is to be expected that the classification suggested above
will require considerable modification when more accurate elevations become available. Although an
attempt to date these surfaces is of very limited value with the scanty data at present available, it is
nevertheless considered worth making as a basis for future discussion. Internal evidence of age appears
to be entirely lacking, and the only feasible approach to the problem is by correlation with adjacent

In the Upper Senegal and Gambia basics, Michel (1959) recognises three principal surfaces.
The highest of these, which varies in elevation from 2,500 feet to 3,200 feet, has been warped and
tilted, and Michel considers that these disturbances took place in the mid-Cretaceous because in the
adjacent Middle Senegal basin the down warped portion of the surface is overlain by Maestrichtian
sediments. Planation of this first surface must therefore have taken place in the early Cretaceious or
late Jurassic. A second surface, which occurs at elevations of 1,500 to 1,800 feet, he considers to be
Eocene. Below this he recognises a fragmentary late Pliocene surface at elevations of 300 to 500 feet,
and three Quaternary terrace levels.

– 29 –
In Sierra Leone, the only concrete evidence yet available is that furnished by the boreholes
drilled through the Bullom sediemtns near Rokell, in the Freetown Peninsula, during 1956 and 1957.
Several holes encountered weathered and lateritised Basement rocks below the sediments at depths
from 150 to 400 feet, and B.H. 9 passed through duricrust from 169 feet (= 150 feet below sea level) to
250 feet. There can be little doubt that the Bullom sediments in this area lie upon a down warped
mature planation surface, and since some of these sediments have been identified as Eocene (Ann. Rep.
Geol. Survey, 1955) the down warping most probably occurred in the late Cretaceous or early Eocene,
planation and prolonged lateritisation having taken place during the Cretaceous.

Correlation of this pre-Eocene surface with on of the diamond field surfaces must be speculative.
Extensive thick duricrust is found only on the Nimini Surface so this would seem to be the obvious
candidate. Going outside the diamond fields, an interesting hypothetical surface can be constructed by
equating the Nimini Surface (2,250 feet) with the Late Sonfon Surface (2,000 feet, also duricrusted)
and the Kangari Hills Surface (1,700 feet). This composite surface would be tilted, with a gradient of
about 16 feet per mile towards the coast and an axis of tilt striking at about 170°. A gradient of this
order would be inadequate to carry the projected surface below the Bullom sediments, and so it
becomes necessary to postulate a considerable increase in the gradient towards the west, indicating
warping rather than uniform tilting. This is compatible with the data produced by the recent marine
seismic surveys of Tennessee Sierra Leone Incorporated, which indicate that the gradient of the off-
short Basement surface is about45 feet per mile.

Wilson (1958) also tentatively correlates the Lake Sonfon Surface with the sub-Bullom surface;
he also mentions the possibility that lower surfaces might be correlated with the sub-Bullom, but this
possibility is excluded here because none of these surfaces show any evidence of the necessary tiling,
and duricrust is thin and scanty upon them. Haggard (1965) has also drawn attention to the tilting of
the more ancient planation surfaces, but he considers the tilting to have taken place in the late Tertiary
or the Quanternary.

Surfaces higher than the Nimini/Sonfon surfaces do of course exist in Sierra Leone, but only as
small remnants flanking and capping the Loma and tingi mountains, which are isolated granite masses
of inselberg type. On these mountains fragmentary surfaces occur at various levels up to 6,000 feet,
but they are not found elsewhere in the country and that author considers it improbable that so little
could remain of Cretaceous or Tertiary surfaces. It is therefore suggested that the high bevels of the
Loma and tingi Mountains represent planations which began in the Palaeozoic or early Mesozoic.

Around the ~Freetown peninsula, there are several features which throw some light on changes
in sea-level. The coastal platforms (“raised beaches”), which almost certainly represent form sea-
levels, were first recognised by Dixey (1919). The information available at present on this subject is
summarised by Wells (1962), who records the four main levels as 750 feet, 450 feet, 80 to 180 feet, and

35 to 75 feet. These probably represent Late Tertiary sea-levels, in which case it would appear that
large parts of the coastal Plain Surface, and possibly also the Tongo Surface, are of marine origin.
Present coastal features, including duricrust-in-place exposed at low tide and various aggradational
features, show that sea-levels are now rising.

To summarise the author’s conclusions:-

1) The Nimini surface is considered to be probably of Cretaceous age by correlation with the sub-
Bullom surface, and therefore to correspond with Michel’s First Surface.

2) The Main Plateau Surface is thus probably Eocene, by analogy with Michel’s Second Surface.

3) The other important surface, the Coastal Plain Surface, resembles the Pliocene Surface of the
Falema/Upper Gambia Plains as described by Michel, and furthermore has the expected
elevations for a surface controlled by the accepted pre-Glacial sea-levels.

The following provisional dating is therefore suggested, with arbitrary dates interpolated for the
intermediate surfaces of whose age there is as yet no indication:-

– 30 –
Age of Planation


I. Nimini (2,250 feet)


II. Main Plateau (1,500 feet)

? Oligocene

IV. Thousand-foot (1,000 feet)

? Miocene

(III. Koidu (1,250 feet)

(V. Tongo (750 feet)


VI. Coastal Plain (100 to 600

feet) 2nd and 3rd High Terraces


VII. Bullom (0 to 50 feet), 1st
High Terrace, Low Terraces

( Rising sea levels. Part-


Submergence of Bullom
( Estuarine aggradation

Although it has been suggested here that a series of surfaces exist in south-east Sierra Leone
which compare with those found in other parts of Africa, it is not implied that the processes of
planation have been identical in operation. Under present climatic conditions in Sierra Leone, inland
scarp retreat over granitic basement rocks takes place almost entirely through the drainage network,
dissecting the topography into compartments. In the final reduction of these compartments, processes
of slope flattening become increasingly important. The possibility that some of the surfaces may be of
marine origin must also be borne in mind.



Diamond deposits in Sierra Leone fall into three categories:- primary, eluvial and alluvial.
Beach and marine deposits have not been found.

The catergories may be defined as follows:-

Primary Deposits – Dimaonds in their original source rock, which, in all proved examples in Sierra
Leone, is either kimberlite or kimberlite breccia. The kimberlite may be either fresh or weathered to
clay, but should still be in a place and undisturbed.

Eluvial Deposits – Diamonds in surface residual deposits (soil A and B horizons, laterite, swamp
gravels) which have been formed by the weathering and selective erosion of underlying or immediately
adjacent primary deposits. Lateral displacement of the diamonds or of the residual medium, where it
has occurred, has been of very limited extent and has not taken place along a drainage channel.

Alluvial Deposits – Diamonds in alluvium, which is material that has been transported and deposited
by running water.

In many cases, an eluvial medium has resulted from the weathering and destruction of
formations which included alluvial diamond deposits. Some or all of the diamonds in such an eluvial
medium have therefore an alluvial history and are correctly referred to as alluvial diamonds.
Generally, the presence of water word corundum or pebbles in an eluvial deposit will indicate the
inclusison of alluvial material but in their absence there is doubt. Normally, however, diamonds in
swamp gravels are classified as alluvial unless positive evidence excists to the contrary. Proved
deposits of exclusively eluvial diamonds are rare, since most kimberlite zones have at some time
formed drainage routes.

– 31 –
It will be apparent from the above discussion that the distinction between alluvial and eluvial
deposits is very blurred, and certainly has no significance from a mining point of view. They are
therefore described together in the descriptive notes which follow, and, throughout this Bulletin, both
are considered to be embraced by the term “Alluvial Diamond Deposits”. Primary deposits are dealt
with separately, in Part IV.


(a) Summary

The types of alluvial deposit which occur are:-

River Channel gravels

Recent flat or Flood Plain gravels

Low Terrace gravels

High Terrace deposits, which can be subdivided into:-

Gravel in place

Disturbed terrace gravel

Gravel residue

Swamp gravels

Penetration deposits

These modes of occurrence are discussed at length in the pages which follow, but in order to
reduce repetitive descriptive matter and facilitate comparison, the characteristic features of each are
summarised in Table 6. *See pages 34 & 35

(b) River Channel Gravels

For the Alluvial Diamond Mining Scheme, channel gravels are the most important single type of
deposit in the diamond fields. Between 50% and 55% of current production comes from the Sewa
channel. About 42% of total past production has also come from there. An estimated 41% of the
country’s remaining diamond resources are in the Sew channel, which is 105 miles long, and the total
quantity of diamonds originally present in this river is estimated to have been 11.4 million carats. The
average width over this distance is 225 yards and the average diamond density, therefore, about 0.27
carats per square yard. The actual density for different reaches showed considerable variation, from

0.71 carats per square yard in the Barma area, to 0.23 near Baoma Oil Mill, and 0.04 in the Wubunge

All rivers in the diamond fields have stepped profiles, which consist of long reaches with
shallow gradient separated by short sections of steep gradient. The steeper sections contain rock-bars,
shallow rapids or narrow deep rock-floored channels, and are generally situated where the river crosses
zones of more resistant rock. In all sections, the channel bedrock surface is irregular in detail, both in
profile and cross-section. Deposition of diamondiferous gravel is controlled by the irregularities,
which are in turn controlled by structural and compositional variations in the granitic bedrock. These
conditions produce extreme variations in gravel thickness and diamond content.

The different situations in which diamondiferous gravel occurs in river channels may
conveniently be grouped in four categories for the purpose of discussion. These are:-

– 32 –
1) Potholes on bar rock surfaces

Potholes usually contain small quantities of medium to fine, well-rounded quartz gravel. The
total quantity of travel is always insignificant from a mining point of view, but they can contain high
concentrations of heavy minerals and can thus be very useful in reconnaissance prospecting. Potholes
occur both above and below the water mark. Those above are on rock-bars which are exposed during
the dry season, and many are barren of heavy minerals. There are always some, however, in which
concentration has taken place, and in the case of the Sewa, recoveries as high as 15 carats per cubic
yard were originally made from a few of the potholes on rock-bars.

Potholes below the low water mark, and therefore permanently submerged, occur in rocky
channels where current velocities are usually high. They nearly always contain good values, although
difficult of access, and in earlier years yielded a high proportion of the diamonds produced by the
thousands of miners engaged in skin-diving in the Sewa. In some of the more rock-bound reaches of
the river, the only gravel present was in such potholes.

2) Boulder beds or natural riffles

These are sections of river bed over which are spread boulders and large cobbles. The spaces
between the boulders are partially filled by coarse gravel, which almost always contains good
concentrations of heavy minerals. In most cases, although the gravel is thin, the underlying bedrock is
decomposed to clay, indicating that the gravel has been slowly accumulating over a long period of
time. Sand overburden is normally absent, and consequently, at times of high water, the deposits must
still be collecting concentrates. These deposits are common in the subsidiary channels of rock-bars,
carrying only a trickle of water in the dry season, and they are consequently important in
reconnaissance prospecting.

3) Pools

These come in all shapes, sizes and depths. They consist of depressions in the river bed, wholly
or partly filled with sediment. The word “pool” refers to the section of deep, slow-moving water which
often indicates the presence of such a bedrock depression. In some pools there is no gravel, and in
some the gravel that is present is barren. Along the River Sewa, however, many hundreds of pools
exist and originally about half of them contained thick payable gravel. Very large reserves of gravel
were once present, as some pools are 200 feet or more in diameter and gravel thicknesses up to 30 feet
have been noted. Values in the region of 5 carats per cubic yard are not uncommon and, occasionally,
pools are found in various sections of the river with values well in excess of this.

About 30% of the known pools in the Sewa have already been mined, chiefly the more
accessible ones. Many large pools remain untouched, however, either because the water is very deep
or because there is thick sand overburden or because the pool occupies most of the channel width and
cannot, in consequence, be isolated by coffer dam.

Valuation of pool deposits prior to mining is at present impossible, since representative samples
of sufficient bulk cannot be obtained except by mining. At the exploration stage, the essential tasks are
to confirm that a pool does in fact contain gravel and that it has not been mined before. The thickness
of the gravels can be determined by the use of platform-mounted Banka Drills, and it is often possible
for experienced divers to sample the upper gravels. Along the Sewa, gravel samples obtained by Banka
drilling can usefully be examined for their corundum content, which will indicate to what extent heavy
mineral concentration has taken place.

In the final analysis, however, expenditure on pool mining is always something of a gamble,
although, in the case of pools in the Sewa, if the presence of virgin gravel has been confirmed then the
odds are most favourable.

– 33 –

Type of
Deposit Occurrence

Nature of the








On beds of major
rivers in diamond
fields, Sewa, Bafi,
Male and Moa Rivers

Coarse, clay-free gravels,
Numerous rounded quartz
pebbles. Usually iron-
stained, sometimes bleached

From 3”



Below the Recent Flat,
a semi-continuous
wide strip of flat
alluvial ground which
occupies lowest part of
river or stream valley.
Scarce along major-
rivers (Sewa, Moa).

Coarse to medium, clay free
quartz gravel. Often
bleached on decomposed
granite bedrock. Continuous
sheet underlies floor-plain,
except over buried rock-

1’3” 4’


As long discontinuous
strips flanking both
major and minor rivers
in diamond fields, but
rarely along streams.

Coarse to medium quartz
gravel. May be clay-rich.
Usually stained, sometimes
cemented, by iron oxides.

1’3” 4’


As short discontinuous
strips along river
valley sides. Also
commonly along old
courses, now
abandoned by river.

Coarse to medium brown-
stained quartz gravel. Often
clay-rich. Varying degrees
of induration: loose and
sandy; clayey with
concretions; hard laterite.

1’6” 8’


Capping low flat-
topped hills and ridges
along river valleys and
old courses. Flanking
lower terraces. This is
dissected First High

Abundant deeply-stained
rounded quartz pebbles in
matrix of lateritic gravel.
Base and top of pebble-
bearing zone not sharply

18” of



As above, and also as
irregular patches on
slopes and crests of
subdued topography,
usually within a few
miles of major river.
May be dissected First
or Second High

Rare stained and corroded
quartz pebbles in surface
loam. Heavy minerals
distributed through a and B
soil horizons. An eluvial
formation containing some
alluvial materials

4’0” 10’

Angular quartz gravels,
continuous sheet below
swamp. Usually bleached.
Coarse to fine, clay-free or
clay-rich. An eluvial
formation usually containing
some alluvial material.

10” 2’


Below swamps, which
form ¾ of drainage
network throughout
diamond fields. High
proportion of swamps
diamondiferous in
certain areas e.g.
Yengema, Tongo,
Lower Sewa.
Elsewhere, more


Below low and high
terrace deposits in
certain parts of
Yengema field.

Granit or schist bedrock
thoroughly decomposed to
gritty clay.

6’0” 10’

– 34 –


Nature of the


Thickness of


Of values



Sand and sandy gravel 5’ Normally in the basal
gravel, but very low values
also present in fine gravel of
the overburden. In pools,
very thick gravels may
occur with good values
throughout and no bottom

Most of payable gravels are
below low-water level. Dry
season water depth varies
from 5’ to 70’. Rock-bar
deposits are exposed in the
dry season.

Mud, clay, silt,
sand, sandy gravel.
Loose and

15’ Highest values in the basal
gravel, with concentration
of diamond at the
bedrock/gravel contact. A
few diamonds in top few
inches of bedrock.

In dry season, water table is
in gravel bed or near base
of overburden. In wet
season, water table rises in
overburden and maya reach
surface. Effluent water
makes pit sides unstable.

Loam, clay, silt, sand,
sandy gravel. Always
semi-consolidated and
free standing because
of incipient

12’ As above

Dry season water table is in
basal gravel, or may fall
below bedrock surface.
During wet season, may
rise to within few feet of

Loam, lateritic clay
with some laterite
nodules, cemented
sands and gravels

6’ As above, but with
increasing proportion of
diamonds in top few inches
of decomposed bedrock

No water in dry season.
During wet season, gravel
may become water-bearing.

Thin sandy loam

9” Values irregularly
distributed through
disturbed alluvial gravel and
underlying 1’ or 2’ of
lateritic gravel.



Usually none, but
occasionally a thin bed
of barren loam or
lateritic clay overlies
the diamond bearing
lateritic gravel in level

nil Values, usually poor,
distributed through thick
zone of lateritic gravel and
decomposed bedrock. May
be concentration at ground
surface if heavy soil erosion
has followed farming.

Organic material, mud,
clay, silt and sand

5’ Values restricted to basal

Water table invariably
above bedrock surface, but
actual level and rate of
discharge into excavations
are extremely variable.

Gravel plus overburden
of either low or high

variable Values usually poor. From
upper surface, increase
steadily to peak at certain
depth, then rapid cut-off
with increasing depth.

If under high terrace,
generally dry. If under low
terrace, water enters from
terrace gravel.

– 35 –
4) The Main Channel Floor

In the permanently submerged parts of the river channels, the deposits which occupy the greatest
area are patchy gravels of variable thickness lying on an irregular bedrock surface and overlain by
sandy overburden. In the case of the Sewa, dry-season water depths average 12 feet, overburden
thickness averages 5 feet and gravel thickness averages 2 feet 6 inches.

About four-fifths of the gravel is non-enriched transit gravel with very poor diamond values,
often about 0.07 carats per cubic yard along the Sewa. This type of gravel may overlie richer basal
gravels or may lie directly on bedrock, which will invariably be hard.

The remaining gravel is a semi-permanent deposit and is enriched in heavy minerals to varying
degrees. In the Sewa, diamond values range from 0.1 to 8.0 carats per cubic yard in this basal enriched
gravel, much of which is found in bedrock depressions and crevices. The underlying bedrock is
normally soft and decomposed, and if it is granite, then the basal gravels may be bleached. The fact
that these basal gravels contain a higher proportion of well-rounded pebbles than the transit gravels,
noted by several observers, has not been satisfactorily explained.

Deposits of this type have been heavily exploited by licensed miners in the past nine years,
principally by diving but also by the construction of coffer dams. Selective mining is practised by
divers, and the average grade of gravel they bring to the surface is about 1.2 carats per cubic yard. In
certain of the richer sections of the Sewa, such as that near Barma, the transit gravels themselves are
often worth washing, if easily extracted, as they sometimes contain values up to 0.2 carats per cubic
yard. A considerable proporation of the payable basal gravel has now been mined from the Sewa, but
enough remains to support diving activity for several years.

(c) Recent Flat Gravels

These are the gravels which occur below the present flood-plain of a river or stream. To mining
companies they are of the greatest interest because values are less erratically distributed than in channel
deposits and the separation of pay from non-pay ground is relatively straightforward from both a
sampling and a mining point of view. Values are generally higher than in terrace deposits, and many
rivers and streams in the diamond fields have flood-plain gravels which contain between 0.3 and 1.0
carats per cubic yard. In most cases, where overburden is not excessive, and provided a sufficient
yardage can be proved, a grade of 0.4 carats per cubic yard will enable mechanised mining to show a

This category includes the bank deposits which are found along the major rivers. ~True flood-
plains are rarely formed along these rivers, for the reasons given in the previous section, but it is usual
to find a narrow strip of alluvial ground, some 30 to 60 feet in width forming the river bank even where
the river is perfectly straight. The mechanism of deposition is not clear, since in most cases the river is
not apparently swinging and opposing banks are frequently both steep and both apparently being
undercut at times of increasing discharge. However, these banks can be flooded in years of
exceptionally high water, and mining operations show that they are underlain by beds of enriched
gravel deposited below the low water level of the parent river and lying on clay bedrock. Values are
usually good, and both from a mining and a genetic point of view there seems to be no reason ti
differentiate these from normal flood-plain deposits.

The exploitation of recent flats presents major problems to the licensed miners and, where
attempted, is often carried out in a very unsatisfactory fashion. The instability of the deep overburden
means that well-designed pit support is necessary, and high water levels result in heavy pumping costs.
Generally speaking, the deeper flats are only tackled by the licensed miners where values are known to
be very high, examples being the flats of the Lower Moinde, the Bafi, and the banks of the Upper
Sewa. Shallow flats, with less than 15 feet of overburden, are exploited throughout the diamond fields
but, in most cases, a flat is regarded as being worked-out after the extraction of about 60% of the pay
gravel containing perhaps 80% of the recoverable diamond reserves.

– 36 –
(d) Low Terrace Gravels

These are older gravels under deep overburden, normally found along the outer margins of
flood-plains or, exceptionally, beside the river channel. They represent earlier flood-plains, recently
incised, and thus resemble current flood-plain deposits in many respects. The important differences
are, firstly, that low terrace bedrock surface is measurably higher than adjacent flood-plain bedrock,
and secondly, that the low terrace overburden, and sometimes also the gravel,, are partially

The partial consolidation of low terrace sediments has been effected by two processes. The first
is the impregnation of sand, silt and gravel with clay minerals, which were carried in suspension by the
river-derived ground waters. The second is the induration of clays by the deposition of lateritic
material. The upper, more clayey, overburden is normally brown and contains soft red laterite nodules.
The lower overburden of clay plus sand, together with the gravel, might be brown as along the Sewa, or
bleached as in the Moa terraces. The overburden thus consolidated is not only self-supporting but is
also of reduced permeability; it is therefore much less difficult for the licensed miners to tackle,
however thick.

The actual elevation of low terraces above the flood-plain varies considerably and is in some
cases very slight at the bedrock surface and nil at the ground surface. The difference is consequently
not always noticeable in a line of pits. However, the distinction is essentially one of age, and
classification in the field is done on the basis of the condition on the overburden (Haggard, 1958). Two
or three low terraces of slightly differing elevations can sometimes be recognised in one section of a
valley. They are usually in the form of discontinuous strips, and correlation is rarely practicable, even
where precise surface levelling is undertaken. Terrace correlation across nick-points and other breaks-
of-gradient is certainly impossible except in a very general way; for example, all low terraces are
assumed to have been formed during the Quaternary. S.L.S.T. engineers classify the lowest and
youngest of the low terraces as “Ancient Flat”, and it is by definition almost at the same level as the
flood-plain, but the validity of this sub-division is questionable.

Values in low terrace gravels tend, on the whole, to be somewhat lower than those in flat
gravels. It is possible that this may sometimes be due to the penetration of diamonds into bedrock, and
that if penetration diamonds were taken into account, and values expressed as carats per square yard,
there would in these instances be no significant difference between flat values and low terrace values.

Low terraces, whose presence was either ignored or unsuspected in the early years of licensed
mining, have become popular with the miners since 1959. Substantial production has resulted,
although not to be compared with that from the Sewa channel. The Sewa low terraces are far greater in
area than the banks and vestigial strips of flood-plain, although they are still of trivial extent compared
to the area occupied by the river channel. Values in exploited deposits fall mostly between 0.25 and

0.8 carats per cubic yard. Unsystematic mining of Sewa low terraces has been particularly wasteful,
the technique of mining in numerous separate pits being especially unsuitable for such deposits with
their deep overburden. As a result, large diamond resources are left in ravaged ground, now


1) High Terrace – Gravel in Place

High terraces are the older terraces, from which part of the alluvium has been stripped by
mechanical erosion, and they are normally found well above flood-plain level. In their typical mode of
occurrence, high terrace deposits mantle the benches, slopes and flat-topped ridges which are found
along the fringes of recent flats and low terraces. They often also occur at some distance from any
river or flood-plain, in which case they indicate the routes of old abandoned river valleys; in such
situations they are frequently unrecognised. The sub-classification of the high terraces which has been
adopted here is based upon the extent to which the original terrace deposits have been disturbed or
removed in any given valley cross-section, the most disturbed and denuded terraces are also the oldest
and highest, but one must be aware of assuming that terraces which are in difference areas but at the
same stage of destruction are necessarily of the same age.

– 37 –
In the first category, gravel in place, the deposits discussed are those in which the gravel bed
remains undisturbed although much of the overburden has been removed. These deposits are dry for
much of the year, but the gravel bed may become an aquifer from July to November, and the water is
then derived from store influent rainwater rather than from an adjacent flood-plain. Overburden is
shallow, and lateritisation is often well advanced. In some cases, the gravel is patchily indurated and
after extraction the more recalcitrant lumps have to be discarded as waste by the licensed miners.

Formation of laterite nodules in the overburden may disguise its identity as an alluvial formation
and lead to the terrace being overlooked during exploration. Another hazard commonly encountered is
the difficultyof differentiating semi-lateritised alluvial clay from semi-lateritised bedrock clay. Terrace
deposits sometimes contain two or more gravel beds separated by beds of sandy clay. When lateritised,
this can be indistinguishable from the gritty clay produced by the decomposition of granitic bedrock,
and mis-identification in these circumstances results in failure to discover the basal gravel, which is
normally the richest. Because of this difficulty, there are partially-mined high terrace deposits
throughout the diamond fields where the basal gravel remains unbroached. No rules can be given for
making a positive identification of a semi-lateritised clay/quartz mixture found beneath a terrace gravel.
When sinking a prospecting or sampling pit, the only safe procedure is to continue sinking until un-
mistable bedrock textures are met.

Values in high terrace gravel in place are generally somewhat lower than those in nearby flat and
low terrace deposits, though once again it is uncertain to what extent the difference is due to loss by
penetration into bedrock. In the deposits which have been exploited so far, values have mostly been
between 0.15 and 0.6 carats per cubic yard. Deposits of this type are ideal for the small mining
company with little capital; provided extraction is systematic and at least 50 cubic yards of gravel can
be treated per day, then an overall grade of 0.25 carats per cubic yard should yield a small profit.

In the exploitation of these deposits by licensed miners there is normally very little waste
because, in the absence of deep overburden and waster, payable ground is almost completely mined out
(excepting the occasional deep undetected gravels mentioned above). Furthermore, as they are able to
exercise their capacity for selective mining to the full, the miners produce appreciable quantities of
diamonds from deposits which would normally be considered un-payable.


2) Disturbed terrace gravel

This is an old terrace deposit from which the overburden has been entirely stripped, the gravel
remaining as the principal constituent of the soil and subsoil. The gravel has not been transported or
reworked, but has been disturbed by the activity of the plants and animals and has become intermixed
with organic material, laterite nodules and rotten bedrock fragments.

Heavy minerals are randomly distributed through the disturbed gravel and the under-lying
lateritic materials which represents the original bedrock. The level of the form bedrock surface cannot
be identified, and it has in any case no significance for the miner. During valuation, each pit must
continue down until values cease. The mining width is generally found to be 3 or 4 feet. Average
recoveries are often disappointing because of the dispersal of the diamonds through several feet of the
sub-surface profile; in most of the deposits examined by the Geological Survey, values have been
between 0.1 and 0.3 carats per cubic yard.

Most of the known deposits of this type have been at least partially exploited by the licensed
miners, who recognise the rounded pebbles (“marble stones”) in the soil, test a few head pans of it, and
being to mine wherever they find corundum. Here again, many diamonds have been produced by the
selective mining of ground with absurdly low overall values. Exploitation of these deposits is often
handicapped by their distance from water, and is therefore usually attempted only during the rains.

– 38 –

3) High Terrace – Gravel Residue

This is the oldest recognisable material of alluvial origin and is widely distributed in the vicinity
of all the major rivers and in many parts of the Coastal Plain Surface. It consists simply of the heavy
mineral residue from a former terrace deposit, all the other components having been either dispersed or
destroyed in the course of terrace dissection. The residue, which consists chiefly of water worn
corundum and the occasional diamond, is distributed through several feet of surface lateritic gravel.
Solitary rounded quartz pebbles may be discovered by diligent search stained and corroded but still

Extensive areas of workable ground are never found, but the licensed miners often search out
small payable patches, and sometimes these are remarkably rich. Much high terrace residue is,
however, virtually barren, and even in highly favourable areas such as the immediate vicinity of the
Sewa River values are usually about 0.01 or 0.02 carats per cubic yard. The deposits are regarded as
consisting of alluvial material in an eluvial medium, but there are many instances where they very
localised high diamond values, unaccompanied by significant alluvial indications, lead to the
conclusion that most of the diamonds are themselves purely eluvial, thus indicating the presence of
local primary sources.

One of the beneficial effects of the Alluvial Diamond Mining Scheme has been the discovery
and exploitation of many gravel residue deposits which would otherwise have remained unknown, to
locate them being very costly in terms of labour. There can be little doubt that many more deposits of
this type remain to be found.

It has been suggested that in certain “high terrace” deposits of largely residual origin, where the
values are present in lateritic gravel, up to 90% of the diamonds may be enclosed by laterite nodules
and can only be recovered by grinding the gravel tailings after washing. It would appear that the
diamonds and other heavy minerals form the nuclei for laterite deposition. If this phenomenon occurs
throughout the diamond fields, then large numbers of gravel residue deposits must have been either
overlooked or seriously undervalued.

(f) Swamp Gravels

Swamps are one of the commonest topographical features of south-east Sierra Leone, as they
constitute 80% of the complete drainage network and about 12% of the total surface area is swamp. A
complete physical description is given in the preceding Section.

Below all swamps is a thin basal layer of gravel, and they have in the past been one of the most
important types of diamond producer, second only to the Sewa channel. They are now of steadily
diminishing importance because few new payable deposits have been discovered in recent years, and it
is the belief of the writer that very few remain to be found as it is now exceptional to cross a swamp in
the diamond fields which does not show evidence of former prospecting. Swamp mining still
continues, but in the majority of cases the miners are either re-working formerly productive ground
where gravel extraction was incomplete or are mining barely payable ground which is adjacent to an
old mined-out rich deposit.

Swamp gravels are eluvial, but they normally contain some material which has an alluvial
history. Many of the productive swamps cut through high terrace ground and in the process of
formation have effected a drastic re-concentration of the heavy minerals. A typical example would be
a s follows:-

A high terrace gravel residue contains average diamond values of

0.01 carats per cubic yard over the uppermost 6 feet of lateritic

– 39 –
Where this old terrace is dissected by a swamp network,
the values from 2,000 square yards of terrace will be concentrated
in the 9 inch thick basal layer of gravel under 1,000 square yards of
swamp. The concentration factor is 16, and the average grade of
the swamp gravel will be 0.16 carats per cubic yard. The
distribution of values will be erratic in details, and there will
be erratic in detail, and there will be some small blocks of gravel
containing 0.2, 0.3, 0.4 and 0.5 carats per cubic yard surrounded
by much near-barren ground. A group of licensed miners could
profitably recover most of the diamonds in this swamp by selective
hand mining, but a mining company would not hope to do so.

In most of the hundreds of swamps that have been exploited during the last ten years, average
grades have been between 0.05 and 0.5 carats per cubic yard, with occasional deposits go9ing well
above these levels. The average deposit grades do not, of course, correspond with the recoveries made
from the selectively-mined and washed gravel, which are rarely less than 0.2 carats per cubic yard.
Although the licensed miners never formally value a deposit, a comparison of mining activity with
verified recoveries shows that the miners’ present assessment of the profitability of swamp ground is as

Less that 0.1 carats per cubic yard – Un payable, never intentionally mined.

0.1 to 0.2 carats per cubic yard – Low grade, yielding only a bare living.

0.2 to 0.6 carats per cubic yard – Payable, yielding surplus over licence and minimum
subsistence costs.

0.6 to 2.0 carats per cubic yard – High grade. Excess attributers attracted.
The miners’ standards change of course. Before 1958, when many good deposits were still
available but the prices realised were much lower, they were not interested in any gravel
containing less than 0.4 carats of clear stones per cubic yard. It is for this reason that much
current mining activity is taking place in ground rejected as un-payable by the early diggers.

Although swap deposits originally formed a substantial proportion of the total alluvial diamond
reserves, few of such deposits could ever have been worked by a large mining company, in view of
their scattered situations, small individual yardages, and generally unattractive grades. Certainly none
of the reserves left in the many partially mined swamps are likely to interest any mining company; this
has been conclusively demonstrated by the results of the Diamond Corporation projects described by
Barber 1961 and 1963).

(g) Penetration Deposits

These consist of decomposed bedrock which does not appear to be genetically related to
diamonds but into which appreciable quantities of alluvial diamonds have penetrated to a depth of
several feet. Their existence in certain localities is proved, but little is known of their extent and
distribution and no satisfactory explanation of the penetration phenomena have been advanced.

Slight penetration of decomposed bedrock by alluvial values is a well-known, if little
understood, phenomenon. In standard valuation and mining procedures, six inches of decomposed
bedrock should always be included in the mining width. Barber (1963) reports the presence of
diamonds in decomposed swamp and stream flat bedrock to depths as great as 18 inches, and noted the
failure of the licensed miners to detect or recover such diamonds. In disturbed and residual high
terrace deposits, diamonds are always recovered from lateritic gravel and semi-lateritised material
which represent the form bedrock of the original alluvial deposit. In this environment, however, some
downward migration of heavy minerals is not unexpected in view of the visible disturbance of the
medium by plant and animal agents.

Diamonds occurring in bedrock in the circumstances described in the above paragraph are
normally classified as part of the overlying alluvial deposit, and the recorded width of pay gravel must
include the diamondiferous bedrock. Instances occur, however, where the diamondiferous bedrock
must be regarded as a separately assessable formation.

– 40 –
These are the deep penetration deposits, where the values in bedrock often exceed those in the
overlying gravel, and they have been found by Sierra Leone Selection Trust beneath terraces of the
Shongo, Oyie, Pokoda and Gbobora streams. In a typical Pokoda example, poor values in the alluvial
gravel are succeeded by even lower values in the immediately underlying bedrock, but they increase
steadily with depth and reach a peak at a depth of 9 feet below bedrock surface. At this depth, the
grade may be as much as 0.5 carats per cubic yard over a width of 1 foot. Below this there is a rapid
decline in values and a cut-off two or three feet further down.

In the example described, the diamond matrix is typical decomposed granite in which occasional
dipping quartz veins may be observed. Apart from the quartz veins, there is no visible trace of any
introduced non-granitic material and no indication of fracturing or shearing. Any remaining doubts
regarding the alluvial history of the diamonds are conclusively dismissed by the fact that the
penetration diamonds are accompanied by water-word corundum. The depth distribution curve of the
diamonds is closely matched by that of the water worn corundum.

Deep penetration deposits have so far only been detected at the localities mentioned earlier, it
will be noted that all these lie on or near to the Oyie-Shongbo fault zone, one of the most important
regional structures. Whether this association is significant or fortuitous is a matter for speculation. If
such deposits are of widespread occurrence throughout the diamond fields, the large diamond reserves
might remain indefinitely undetected. However, deep bedrock tests carried out by the Geological
Survey below diamondiferous terraces and flats elsewhere in the fields have all failed to recover
diamonds or any other alluvial concentrates from depths greater than 24 inches. The opinion now held
by the author is that deep penetration results from some combination of conditions that will only rarely
exist together, in one locality, and that quantities of diamonds locked in penetration deposits will
therefore be relatively unimportant.


(a) Summary

Alluvial concentrations of diamonds are always accompanied by concentrations of other heavy
minerals. In the course of prospecting operations, heavy mineral concentrates are always collected and
examined, because they afford much useful information and the correct interpretation of their
distribution is an indispensable part of diamond exploration.

In the course of the present survey, concentrates have been collected from every mining and
prospecting site in south-east Sierra Leone, and a large number of mineral varieties have been recorded.
Only some of these varieties are significant of course. The minerals found have been classified as
follows, in accordance with their relationship to diamond occurrences:-

Those with alluvial significance are; corundum, tourmaline, epidot-with-quartz.
Those with genetic significance are; magnesian ilmenite, pyrope, chrom diopside.
One with an apparent relationship, but whose exact significant has not yet been determined, is
Those with no known significance are; zircon, matamict zircon, rutile, magnetite, normal ilmenite,
laterite (goethite and limonite), epidote, columbite, ilmenorutile, haematite, chrysoberyl, spessarite,
almandine, topaz, tremolite, actinolite, pyrite, green spinel, kyanite, cassiterite, gold, vttrocrasie.
The mode of occurrence and significance of these various minerals are discussed in the following

(b) Minerals with alluvial significance

Corundum This mineral is regarded by all licensed miners as an almost infallible indicator
of diamond, and there is a great deal of justification for this. The specific gravity of corundum is very
close to that of diamond, and its durability is second only to that of diamond. Consequently, where
both corundum and diamond are part of the same alluvial system, they tend to concentrate together.
The corundum which has been exposed to the fluvial processes will, however, be abraded to a degree
proportionate to the distance it has been transported.

– 41 –
Angular or euhedral corundum has little significance, as it must be of very local origin and
concentrations therefore reflect proximity and yield of the source rather than the operation of alluvial

In most alluvial deposits, the approximate diamond; corundum ratio can be determined and this,
if used with due caution, can be of assistance in evaluating parts of the deposit from which only
inadequate samples can be obtained or in assessing the significance of abnormally high values.

Water worn corundum has considerable significance when found in diamondiferous lateritic
gravel or swamp gravel. Both these formations are of residual origin and neither has formed by
processes which could have caused abrasion of the corundum. Any water worn corundum present must
have been derived from fluvial deposits, either terraces or an ancient channel, and there is then a strong
presumption that any associated diamonds are alluvial.

In their prospecting, the licensed miners proceed on the assumption that wherever corundum is
found, diamonds will be found. The farms and swamps are searched for corundum, and wherever
appreciable quantities of it are found, large gravel samples are extracted and washed. Many
discoveries have been made by this method, and failures are attributed to supernatural intervention.
The method is obviously valid under certain circumstances. In the vicinity of known diamond-bearing
rivers such as the Sewa or the Matemu, terrace residues are likely to contain diamonds, therefore a
prospecting method which provides firstly for the identification of such residues and secondly for the
location of concentration points within the residue has much to recommend it. However, the method
has serious limitations which are not yet fully appreciated by the miners, and these are:-

1) Only water worn corundum has significance. Angular corundum should be ignored.

2) Corundum occurs universally throughout the diamond fields. The gravels of many rivers are
rich in corundum but carry no diamonds. The resting of old terraces of such rivers is pointless.

3) Primary deposits, or eluvial deposits which have no alluvial component, will be overlooked as
there will be no corundum.

The corundum found in alluvial concentrates occurs in a wide variety of colours; white, grey,
brown, yellow, black, blue, red purple, colourless. Grain size varies from 2mm to 8cm. but
pieces large than `15mm are rare. None of the corundum is clear, in several hundred corundum
concentrate collections, nothing of gem quality has been seen. The source of the corundum has
not yet been established. It is widely assumed to have come from pegmatites, but it has never
actually been found in the rock, and its occurrence in the granites would be distinctly unusual.

Tourmaline Small well-rounded grains of fibrous black tourmaline are abundant in
concentrates from Sewa gravels. As the specific gravity of these grains is only about 3.0 they only
function as concentration indicators in a very rudimentary fashion. Their presence in swamp gravels is
often taken as evidence of the presence of Sewa terrace residues, but this is an unwarranted assumption
as both angular and water worm tourmaline have now been found to be of widespread occurrence on
the Coastal Plain Surface south of Bo.

The source of the tourmaline is unknown. The grains are often referred to as “tourmaline schist”
because of their fibrous well-cleaved structure, and are sometimes attributed to formations in the
Nimini Hills Schist Belt. To the best of the author’s knowledge, however, no outcrops of tourmaline
schist have been recorded from this district.

Epidonte-with-quartz Epidote may be seen in most of the granitic rocks in the vicinity of faults and
shear zones. The Sewa gravels contain a few small greenish pebbles of an exceptionally resistant, fine-
grained rock which consists of inter grown quartz and epidote. The rock is apparently a cataclasite and
must derive from the major fault zones which control the Sewa course. These pebbles function well as
concdentration indicators, and a Sewa gravel which is rich in these pebbles is almost always also rich in

– 42 –
(c) Minerals with genetic significance

The minerals discussed under this heading are known to occur almost exclusively in kimberlites,
and their presence in alluvial concentrates can therefore generally be taken to indicate the proximity of
kimberlite outcrops. In Sierra Leone, quantitative sampling of soils and alluvium for these minerals
has been used to trace kimberlite outcrops in known diamondiferous ground, and in other territories the
method has been used as the basic method of diamond prospecting.

Magnesian ilmenite (Picro-ilmenite). This occurs both in kimberlite and alluvials as well-rounded
grains. Size ranges vary between one kimberlite and another, but even where the smallest sizes occur
very few grains will pass a ½mm screen. Shapes vary, but there is a tendency for the grains to occur as
slightly flattened ovoids, and this tendency increases with decreasing size. In the field, the mineral is
identified by its rounded shape, black streak, conchoidal fracture, and the brilliantly lustrous
appearance of fractured surfaces. In the laboratory, identification may be confirmed by checking the
magnesia content, which should be in excess of 8%. Magnesian ilmenite from the Koidu kimberlites is
usually noticeable magnetic, and grantham and Allen (1960) have shown that this is due to exsolved
magnetite within the grains. Ilmenite from the Tongo and Panguma kimberlites is only very weakly

Larger grains of magnesian ilmenite (i.e. +4mm) are found only within a few hundred yards of
their source, partly because of their weight but principally because of the readiness with which they
disintegrate when subjected to fluvial processes. The smaller grains may travel some distance down
the drainage system, being considerably more resistant to mechanical attack, but it has not yet been
established how far they will travel and still remain recognisable. The effects of solution must be

Ilmenite recovered directly from weathered kimberlite often has a grey appearance caused by a
thin skin of leucoxene, whereas that from alluvials generally has a faintly lustrous black matt surface.
It is often asserted that the distance of the source can be estimated from the amount of surviving
leucoxene, but this is not so. A heavy leucoxene skin certainly indicates proximity of the source but
its absence indicates nothing. At Panguma, weathered kimberlite yields leucoxene-free ilmenite.
Furthermore, it is not possible in the field to distinguish with certainty an ex-kimberlite matt surface,
the result of abrasion during intrusion, from a fluvially-abraded surface.

Pyrope garnet this mineral is less common than ilmenite in the Koidu kimberlites, and is rather
uncommon in the Tongo kimberlites. In the kimberlite, it occurs as slightly flattened ovoid nodules
which usually have a kelyphite skin and sometimes have kelyphite-lined fractures. As the nodules are
well-fractured, they disintegrate readily in stream channels and are usually therefore found as angular
fragments in alluvial gravels. Kimberlite garnet fragments can, however, generally be identified as
such by the recognition of remnants of the original nodule surfaces, which have a characteristic
botryoidal appearance.

Three varieties of pyrope occur in concentrates obtained from the Koidu kimberlites and are
distinguished by their colours, which are dark red (tinged with brown), pale orange-red and pale purple.
The dark red variety is referred to as “common red”; the other two are both uncommon.

The small fragments which result from the disintegration of the pyrope nodules are unfractured
and physically stable, and may be carried for some distance in the drainage. They are especially
abundant in the gravels of the Meya-Moinde system, which drains the Koidu kimberlite zone. The
identification of garnet fragments in an alluvial concentrate presents no difficulty, but with increasing
distance from the source, grain surfaces become modified by abrasion and corrosion, and it becomes
progressively more difficult in the field to differentiate pyrope from the non-kimberlitic garnets which
occur in many Sierra Leone alluvials. Laboratory identification is often necessary, and this is done by
confirming the high magnesia and chrome content of the pyrope grains.

– 43 –
(d) Other minerals

Chromite occurs in most kimberlites in Sierra Leone, but it is not normally considered of value as an
indicator mineral since it is common in other ultrabasic rocks. It has been noted, however, that in
certain sections of the diamond fields, the distribution of alluvial chromite corresponds closely with
that of diamond. Investigation of this relationship has produced only negative results, indicating that it
is probably fortuitous.


Minerals of no known significance

Coarse concentrates in most parts of the diamond fields consist predominantly of corundum,
which has already been discussed. Other minerals may be locally important; in some swamps, for
example, the bulk of the coarse concentrates may be either ilmenite, metamict zircon, rutile or
ilmenorutile. The other minterals which were listed in the summary occur sporadically, and none of
them individually ever constitutes more than 5% by volume of a course concentrate.

Fine concentrates (i.e. those with grain size of less than 2mm), contain very little corundum. In
swamp and minor stream concentrates, either ilmenite, zircon or magnetite always predominates. In
river gravels, over 95% of the fine concentrate is normally an ilmenite/zircon mixture, only isolated
grains of other minerals being present.

Some of the minerals recorded, rutile, gold, cassiterite and columbite for example, would be of
economic interest themselves if present in suitable concentrations and quantities, but no such
concentrations have been seen during the present survey..

– 44 –


For the purposes of description and estimation of resources, the alluvial diamond fields have
been divided into a series of blocks. Within the fields the distribution of deposits is uneven and they
tend to a certain extent to occur in groups; the block boundaries have been chosen as far as possible to
enclose these natural groups.

When reading the following sections, it will be necessary from time to time to refer to the
appropriate sheet of the 1:50,000 diamond Series, because bracketed numbers in the text correspond
with encircled numbers on the maps. Certain blocks are not covered by the encircled number is placed
on or beside a deposit which has been examined by the Geological survey and from which, in the
majority of cases, a gravel sample of about 5 cubic yards has been washed. Deposits without numbers
have been visited, but not sampled or examined in detail. In addition to the thirty-one blocks, there are
also a number of small isolated deposits which could not conveniently be included in any block, and
each of these has been separately described.

The two Sierra Leone Selection Trust Mining Leases have not been included in the system of
blocks, but the deposits within them have been divided into groups for convenience in description.
Only a minority of the S.L.S.T. deposits have actually been visited by the author; nearly all of the
information given here was obtained from Company records, to which free access was given. Precise
locations, which would be of material assistance to illicit miners, have been excluded from all
descriptions of S.L.S.T. alluvial reserves.

The descriptions which follow commence with Yengema Field, the most important alluvial field,
and then proceed in a natural sequence down the Sew Valley to the Sumbuya area. Attention is then
transferred to Tongo field, the description of which is followed by an account of the fields to the south
of it. Finally, there are brief notes on various isolated diamond discoveries which may be of interest for
future exploration.


(i) THE KIODU AREA. (Map Sheet 59)

Within this area are all the deposits of the Meya/Moinde drainage above the confluence of the
two streams. The Koidu area was without doubt the richest alluvial diamond district in Sierra Leone
and probably in the world, and about 9 million carats have been mined from the alluvials within a
radius of 1½ miles of Koidu town. The three principal deposits of this area were the flats of the Woyie
stream, the flats of the Meya stream, and the Wongoyie swamp with its tributaries.

In the flats of the lower and middle Woyie a 4th order stream of low gradient over 200,000 cubic
yards of gravel under shallow overburden contained average values of over 5 carats per cubic yard
throughout. Large blocks of ground gave recoveries of 20 carats per cubic yard or more, and
recoveries up to 77 carats per cubic yard were recorded from some of the prospecting pits. During
mining operations many large diamonds were recovered from the Woyie gravels, the largest being the
770 carat clear stone found in 1945.

The Wongoyle swamp and its tributaries drain the outcrop areas of the principal kimberlite
bodies, and its shallow gravels contained, in some sections, even higher values than those of the Woyie,
although the total gravel yardage was much less. The highest recorded recovery in Sierra Leone came
from a prospecting pit in the Wongoyie which yielded 265 carats per cubic yard. Mining of the Woyie
and the wongoyie deposits was completed some years ago, and their valleys have now been partially
flooded to form a reservoir.

– 45 –
The Meya is a well-graded 5th order stream which flows from south to north across all the
kimberlite dyke zones and receives numerous diamond-bearing swamps and streams, including the
Woyie as tributaries. Wide flats and low terraces extend along the Meya valley, with over burden up to

20 feet in thickness covering the thick diamond-bearing gravels. Gravel yardages were originally
substantial and most of the ground carried values between 1 and 5 carats per cubic yard. This deposit
has for many years been the mainstay of the Yengema production, and about one-third of it still
remains in reserve.

The Meya and the Woyie valleys were flanked by low terrace deposits, although in the case of
the Woyie these were narrow and fragmentary, but it is remarkable that there are no true alluvial high
terrace deposits in the area. There are several blocks of diamondiferous ground classified as high
terrace for mining purposes, but these are in fact eluvial deposits overlying or close to kimberlite

Another major stream in the Koidu area is the Moinde, which is joined by the Meya near
Penduma. Above this confluence, the middle Moinde is a 5th, order stream and its deep flats, now in
process of valuation, are believed to contain payable values as far as Boma. In addition to the principal
deposits described above, most of the smaller streams and swamps within two miles of Koidu carried
diamonds, with values varying over a very wide range, many of them being completely un-payable.

Stimulated by the high alluvial concentrations, the Company has been carrying out intensive
kimberlite investigations in the Koidu area since 1947. It has been established that several kimberlite
dyke zones, carrying good values at the surface, cross the Koidu area south of the Woyie. Three small
pipes of kimberlite breccia are associated with these zones, and underground exploration of these has
now begun. More dyke zones and a fourth small pipe occur to the north of the Woyie, but these appear
at present to have only low values.

The kimberlites are discussed in detail in Part IV but there are some interesting aspects which
could usefully be mentioned here. The tykes vary in width from a few inches to about three feet,
consequently although the total length of dyke outcrop is about 19,500 feet, the total outcrop area is not
large. The pipes also are not large, the two largest each having an outcrop area of about 30,000 square
feet. Calculations made by the S.L.S.T. Prospecting Department, based on a thorough sampling
programme, show that the exposed kimberlites in the Koidu area contain an estimated 2700 carats per
foot of depth.

Dr. Lovering’s results (see Part V, Section A, iii) indicate that the kimberlite bodies are almost
certainly post-Triassic, and it is therefore unlikely that the land surface has been lowered by more than
2000 feet since intrusion. In fact, if the Nimini Surface is as old as is suggested here (Part II, Section
B, ix) then little more than 1000 feet have been removed, this being the difference in present elevation
between the Koidu and the Nimini surfaces. By assuming that the eroded pipe sections were enlarged
and contained enhanced values, it is just possible to concede that the known Koidu kimberlites might
have produced all the diamonds in the Koidu area, which, if all un-payable deposits are included, must
have been at least 10 million carats in aggregate. It is not, hoever, possible to agree with the main
premise of the alluvial distribution theory that has enjoyed general support until recently, that is, that
these kimberlites produced most of the alluvial diamonds in the rest of Yengema field, the Bafi, and the
Sew Valley, a group of fields whose total resources are estimated originally to have been 52 million

(ii) THE YENGEMA AREA. Map Sheet 58)

This area, centred roughly on the mine headquarters at Yengema village, is defined to include
only the Bandafaiyi stream and all its tributaries. Over half of the total drainage network in this area
was diamond-bearing, and about one=-third, including all the principal streams, contained payable
gravels. Values in these gravels ran generally between 0.3 and 1.5 carats per cubic yard, but were often
higher than this in the flats of the Oyie and the Gaiya, which contained a high proportion of the total

– 46 –
The principal producers in past years were the Oyie and Gaiya streams. The Oyie flats contained good
values almost the whole way from the source of the confluence with the Bandafaiyi; these flats are now
largely mined out, but some reserves remain in the low terraces. The flats of the lower Gaiya are also
mined out, and mining continues now in the upper Gaiya, although reserves there are rather slim.
Along the Bandafaiya stream itself there has as yet been little mining, and the flats still contain
substantial reserves of payable gravel.

The Yengema area is similar in extent, density of diamondiferous drainage, and pay gravel
yardages to the Koidu area, but average values in the deposits were on the whole much lower and the
total reserves therefore considerably less. Occasional large diamonds have been found in various
deposits, the largest being a bort stone of 783 carats from the upper Gaiya.

Kimberlite exploration in the Yengema area has been less intensive than at Koidu, but several
kimberlite dykes have been encountered in the course of mining operations. The main kimberlite zones
of the Koidu area, designated zones A to G on the mine plans, continue across the Yengema area with
the same approximate strike of 67° true. The actual dykes on each zone are, however, short and
discontinuous and diamond values appear to diminish from east to west. In addition to the normal
kimberlite zones, two small kimberlite dykes have been discovered beside the Oyie near Bumpe.
These contain high values, but are also abnormal and of exceptional interest in that their strike is about
12° ture, which is parallel to that of the large Oyie-Shongbo fault zone on which they are situated.
Grantham (1960) has drawn attention to the fact that none of the kimberlite dyke zones can be traced to
the west of this fault. As there are no exposures, dyke-fault relationships at the points of intersection
are obscure and urgently await clarification. This fault zone is also remarkable for the occurrence
along or near it of the puzzling deep penetration deposits, which are not known to occur elsewhere.

It is generally recognised that the few known kimberlite dykes in the Yengema area are
inadequate to account for all the diamonds in the alluvials. The explanation which has been widely
accepted in the past is that a large proportion of the Yengema diamonds were carried there from the
Koidu area by an ancient large river, and there is of course some supporting evidence for this. Both
Koidu and Yengema are basin-like areas of moderately flat country and they are connected by the
“Motema Gap”, furthermore many of the stream gravels carry quartz pebbles whose degree of
roundness is inconsistent with the stream order. The theory of an ancient river was first put forward in
1935 by A.C. Clarke (Pollett, 1937) who also, however, drew attention to the principal objection to it,
that is, the markedly differing colour and shape characteristics of diamonds from different alluvial
deposits. He pointed out that these differences could not have resulted from fluvial concentration
processes and that a common source is therefore excluded. It has been possible, by comparing output
of different Pan Plants, to confirm that these differences do exist, and Clarke’s objection is therefore
regarded as conclusive.

It follows that the greater part of the alluvial diamonds in the Yengema area must be attributed to
local sources, ancient drainage being responsible only for local redistribution and concentration, and it
is therefore to be expected that further exploration will uncover many additional kimberlites.

(iii) THE LOWER MOINDE (Map Sheet 58)

Only the Moinde river and its tributaries, from the Moinde-Meya confluence downstreatm to the
northern Mining Lease boundary, are included within this area. The Lower Moinde is a 6th order
stream whose valley is narrow and straight, its alignment being rigidly controlled by the Moinde Falut,
and it seems unlikely that the drainage here has ever deviated appreciably from its present route. The
mean gradient of the valley floor is only 6 feet per mile, however, consequently the course of the actual
river channel is tortuous in detail, and the valley floor contains numerous graded sections where gravel
has accumulated and flood-plain has formed.

These flats, from which production has only recently begun, contain substantial diamond
reserves which probably form the most important section of the remaining alluvial reserves of
Yengema field. The pay gravel in the flats lies below deep unstable over burden with a continuously
high water table, and this has so far protected it from the illicit miners. There were also, originally,
several low terrace deposits and some strips of high terrace, all of which contain good values. These,
however, have all been laid waste by illicit miners and could not now be worked at a profit.

– 47 –
Payable gravels also occur in the small streams of the Komau area, associated with a kimberlite
dyke, but apart from these there are only a few small scattered low-grade deposits in Lower Moinde
tributaries. It is generally assumed that the Lower Moinde diamonds have come principally from the
Koidu kimberlites, and this assumption could be essentially correct, as these kimberlites are only a few
miles away and are all drained by the Meya, the principal Moinde tributary, following a drainage route
of long standing.


This area covers a three-mile section of the Bafi river to the south and east of Yomadu, and
various Bafi tributaries. There is no flood-plain along this section of the Bafi, but the channel gravels
carry diamonds and there are strips of high terrace and low terrace.

In the Bafi channel, no reliable method of sampling the channel gravels has yet been devised,
but during prospecting operations good values were recovered from rock-bar deposits, and there is a
high incidence of illicit mining by divers. Whether the bulk of the Bafi diamonds can be profitably
recovered is at present open to question. Sierra Leone Selection Trust installed an experimental jet-lift
dredge in 1962 but although this successfully mined and treated the river channel deposits, average
diamond recoveries were too low to cover operating costs and mining ceased in 1964. Many of the
minor Bafi tributaries contained moderate diamond values but all those on the north bank, together with
the Bafi tributaries contained moderate diamond values but all those north bank, together with the Bafi
terraces, have been stripped by illicit miners, who have a convenient base at Yomadu.

The major exploitable deposits of the area are those of the Shongbo River and its tributaries.
The gravels of the Shongbo flats carried diamonds throughout, but much of the middle Shongbo is now
mined out, recoveries having been mostly between 0.5 and 1.5 carats per cubic yard. Reserves remain
in the Upper and Lower Shongbo; the Upper Shongbo contains only small reserves, whilst the Lower
Shongbo deposits have not yet been closely sampled and the available data indicate that they may be
unworkable because of low average values.

Heavy minerals of kimberlitic origin are rare in this area and there are no known kimberlite
outcrops, although the Komau zone should cross the Shongbo headwaters. Large diamonds are also
rare. The Shongbo diamonds have, in the past, been attributed to the Koidu source area, and it has been
supposed that they were transported thence by the hypothetical early drainage, which is assumed to
have swung northwards after passing Yengema. There is of course, no evidence whatsoever to support
all these suppositions, and until there is, the simpler explanation must be preferred and these diamonds
attributed to a local source, one deficient in the usual heavy minerals.


This area contains the Gbobora stream from the point where it enters the Mining Lease to its
confluence with the Bafi, and a section of the Bafi River extending three miles upstream and three
miles downstream from the Gbobora confluence. In 1930, the first Sierra Leone diamond was
discovered in this area, at Fotingaia on the Gbobora, which was a most surprising site for the discovery
in view of the fact that the deposits in the vicinity are of very low grade and unlikely ever to be mined.


channel bed consists of bar rock from which all sediment has been scoured. Although numerous
sediment traps doubtless exist, major gravel accumulations will be present only in a few of the deeper
pools and no satisfactory method of sampling these has yet been developed.

Only vestigial terraces and flats can be found along this section of the Sewa, and tributary
deposits are limited to a few small swamps which are near to the river and occupy recently abandoned
channels. Most of these tributaries contain only small yardages of low grade gravel, and those few that
did contain some payable gravel have already been plundered by the illicit miners.

In summary it can be said that, on the basis of the limited information available, this appears to
be one of the sections of the Sewa with the least potential. The diamonds that are present probably
came from the Tefea area. It is interesting to note that there are a few minor deposits in the small area
to the north of Gbambiadu Falls which is enclosed by the Bagbe/Sewa. Heavy illicit mining of swamp
and terrace gravels has persisted for some years, in the face of repeated attempts at suppression. These
deposits are related to old courses of the Bagbe River and, since the Bagbe is not known to carry
diamonds, it has become customary to assume that the diamond-bearing Bafi River formerly joined the
Bagbe above Bagbema. There is no evidence either for or against this assumption, but if it should
happen not to be correct then a diamond source must be present in the Lower-Bagbe.


In 34 years of mining, Sierra Leone Selection Trust has produced about 18 million carats of
diamond from Yengema field, whilst losses by illicit mining and other forms of theft are estimated by
the author to have been in the region of 2 million carats. The company’s formal reserves of payable
gravel, on the 31st December 1965, contained 2,488,480 carats. The author estimates gross remaining
alluvial diamond resources in the Field at 7½ million carats, but this figure includes not only reserves
of all categories but also the substantial quantities of diamonds which are dispersed in un-payable
ground. Alluvial mining in the field is thus seen to have a limited life, the actual length of which may
be anything from ten to fifteen years, depending largely on the extent to which illicit mining can be

The long term future of the field depends on the underground exploitation of kimberlite bodies.
The total resources of diamonds in kimberlite are likely to be enormous, but only a fraction will be
present in bodies which can be mined at a profit. In fact, at the time of writing, it has not been
established whether any of the known kimberlite bodies are payable. As will have been apparent from
the preceding pages, the author maintains that many more kimberlites or other source rocks remain to
be undiscovered.

Looking at Yengema field as a whole, one feature which stands out is the close relationship
between diamond distribution and the structural pattern, made strikingly apparent by the triangle
formed of the Moinde Fault (340°), the Oyie-Shongbo Fault (13°), and the main kimberlite dyke zones


Block 1 – PEYIMA. (Map Sheet 47 and 58)

The Peyima area was for many years the scene of exceptionally intensive mining activity, and
total production to date is estimate at 856,000 carats. The principal deposits were those of the Bafi and
Lower Moinde rivers.

The flats and low terraces of the Lower Moinde (1) contained gravel of uniformly high grade
below 15 feet to 25 feet of over burden. These deposits are now virtually worked out, but they were
sampled in 1962 by the Geological Survey, and the average grade was then estimated to be at least 2.0
carats per cubic yard. The gravel included a proportion of well rounded quartz pebbles, but this feature
is consistent with the stream order, which is 6th. The heavy mineral concentrates were not particularly
abundant, about 15 c.c. per cubic yard, and contained chrysoberyl, tourmaline, epidote, magnesian
ilmentie, pyrope and unidentified round grey concretions in addition to the ubiquitous corundum,
zircon and normal ilmenite. The median size of the diamonds recovered was 0.45 carats, about 405 of
them being coated stones, and half of the clear stones were of poor quality with numerous inclusions.
The Lower Moinde diamonds are generally assumed to have been transported directly from the Koidu
area by the Moinde-Meya drainage, a few possibly having been added by minor kimberlite zones
crossing the Moinde valley, and there are at present no grounds for questioning this assumption.
The deposits of the Bafi River (2) appear to be a direct downstream continuation of the Moinde
deposits, since the diamonds have similar characteristics, and values cut off abruptly in the Bafi above
the confluence. The Bafi channel was formerly very productive, mining having been carried out for
many years by diving and by the construction of coffer dams, but most of the accessible gravel has now
been removed and there is little activity at present. Fairly substantial reserves remain in the deeper
parts of the channel and are suitable for exploitation by a small mining company. In 1961 and 1962, an
attempt was made by a Native Company to drain the Bafi loop to the north-east of Dandiadu by
diverting the river through a channel cut across the neck of the loop. Unfortunately, the excavation of
the channel by dragline was never completed and the project was abandoned, apparently through
exhaustion of capital.

This section of the Bafi has almost continuous flats and low terrace deposits (2 and 3) from
which about half the gravel has now been mined. Some of the remaining reserves have been
disregarded because of deep unstable overburden, but the greater part exist as remnants and pillars
between old paddocks. Average values were good, about 0.72 carats per cubic yard, and the basal
gravel was usually 18 inches to 24 inches thick beneath about 18 feet of overburden. In some places,
however, the gravel is absent, the overburden lying directly on hard rock, and this has been taken into
account when calculating the average gravel thickness. High terrace deposits have been mined at
Dandiadu (4) and north of Peyima, but in neither of these areas have values been particularly high.
Heavy mineral concentrates along the Bafi are similar to those of the Lower Moinde, but magnesian
ilmenite and pyrope are very rate in the +1mm faction, although reasonably common in the -1mm

An interesting small deposit is that along the Peyi stream beside Peymia town (5). Here there is
a wide stream flat and blanking low terrace, with a maximum aggregate width of 800 feet. The
overburden is from 9 feet to 12 feet in thickness, and the gravel contains a few small well-rounded
quartz pebbles. All these facts are inconsistent with the size of the present stream, which is only 4th
order, and indicate that the Peyi follows the old course of a larger stream. An inspection of the
topography shows that this might have been either the Moinde or the Shongbo, both of which carry
diamonds, but whichever one is chosen, it then becomes necessary to account for the barrenness of the
terrain between its present course and the Peyi deposits. Values in the Peyi were not particularly high
but the deposit has been quite intensively mined because the stones recovered had an exceptionally
high median size and consequently realised high prcies per carat. ~From the inadequate data now
available, the median size is estimated to have been about 2¼ carats. A trace of magnesian ilmenite
occurs in the Peyi concentrates but no pyrope has been observed.

Scattered low values are present in the Yifi swamp (6), which could have been an occasional
alternative course of the ancient drainage through the Peyi. Between Peyima and Sukudu are two small
swamps (7) which are now completely mined out which originally contained high values. The swamp
gravels consisted of re-worked Moinde terrace material, yet the stones recovered appear to have had a
median size of about 2 carats, which is far above the median size in the nearby Moide flats.

The small swamp deposits near Moindema and Wauda (8 and 9) do not appear to be related to
old Moinde courses or terraces, as the gravels contain neither rounded quartz pebbles nor water worn
corundum. The concentrates do contain rare grains of manesian ilmenite and fragments of a dark
reddish-brown garnet, often euthedral when under 2mm in size.

Block 2 – THE MIDDLE BAFI (Map Sheets 47 and 58)

All activity in this Block is based on the town of Yomadu which lies not within the Block but
inside the S.L.S.T. Mining Lease. The main diamond producer has been the Bafi river and its low
terraces and flats, from the Lease boundary to Hemakono. Intensive mining began illicitly in this area
in 1953, and contained under licence after the introduction of the Alluvial diamond Mining Scheme in
Beginning from the Lease boundary, the first one-and-a-half miles of Bafi channel (11) may
justly be described as the most intensively mined section of river channel in Sierra Leone. Here the
techniques of aqualung diving were first introduced by E.I. Bragg. Recoveries of 10 to 15 carats per
cubic yard were made from gravel selectively mined by Bragg from natural channel-floor riffles on
exposed hard schist bedrock, and good recoveries were common throughout this section. Very little
gravel remains now, except in the centre of the large deep pool opposite the Kamayi confluence. The
margins of this pool have been mined by coffer dams but, at the time of the author’s last visit in 1965,
the centre had not been tackled.

From the Sumunji confluence down to the Hemakono area extraction of gravel has been less
complete, and the character of the channel changes somewhat. Many sections of the channel are
narrow, with a floor of smooth hard rock and little sediment accumulation. Numerous small pools exist
in the wider sections, however, and good recoveries have been made from most of these. Near
Hemakono, the river crosses the eastern margin of the main Nimini Hills schist Belt and enters a
narrow gorge. Two pools with water depths of 85 feet and 100 feet occur at the entrance to this gorge
and were successfully mined by Bragg in 1964. Downstream from this point, there has been very little
mining, partly no doubt because of the relative inaccessibility of the area and also because of the
generally deep water. Possibly, also values may be poor. Certainly it can be said that many interesting
pools remain untouched.

Flats and low terraces occur only in the first 1½ mile section of the river (12), and are usually
less than 200 feet in width. These have been very intensively mined in the last five years, consequently
less than half of the original gravel now remains. In the widest part of the low terrace, opposite the
Singapanda mouth, exceptionally good values were recovered from some of the paddocks and the
gravel from one yielded 12 carats per cubic yard.

There are various swamp and high terrace deposits (13 to 19) in this Block, and all appear to
represent old courses of the Bafi river. The most important deposits were the Singapanda. The high
terrace deposit consisted of a disturbed gravel which the miners had completely stripped down to hard
lightly decomposed granite, long before Geological Survey examination. Consequently the estimate of
recoveries is very approximate indeed, based only on the examination of tailings and verbal reports.

The Singapanda swamp (14) contained a normal swamp gravel with high average values of
about 1.6 carats per cubic yard and was mined repeatedly for several years. It is now regarded as
mined out, but diamonds still remain in tailings, small pillars between old pits, and top bedrock; a few
miners can usually be found making small recoveries from these sources. The presence of water worn
corundum and a few pebbles in the gravel show that this is an old river course which must be presumed
to have been occupied by the Bafi river although, from an inspection of the topography, the Kamayi
would appear to be equally eligible.

Diamond characteristics in the middle Bafi differ slightly but no decisively from those in the
Peyima area. In the few parcels seen by the author, the median size was somewhat higher and the clear
stones were in general of better quality. All the diamondiferous gravels in this Block contain small
quantities of kimberlitic heavy minerals in the concentrates. It is probable that the diamonds here have
come partly from the Koidu area and partly from sources in the Shongbo area, and that the relatively
high alluvial concentrations are caused by the frequency of schist bodies in the bedrock. These bodies
intersect both present and ancient drainage routes at an angle. The possibility of a local source cannot
be altogether excluded, however. Some of the swamp deposits, for example that in the Kamayi
tributary (17) and that in the north fork of the Singapanda, are a little oddly situated in relation to old
Bafi courses.

Block 3 – NIMIKORO. (May Sheet 58).

This Block contains one major deposit, that of the Gbobora below Mjaiama Nimikoro, together
with a number of minor deposits from which there has been only insignificant production. They are all
of unusual interest, however, in that although they are adjacent to the Yengema area, the diamonds are
unquestionably distinctive and must be attributed to local kimberlites, which have not yet been found.
For a distance of about 1¼ miles upstream from the S.L.S.T. Lease boundary, the Gbobora flats

(22) are about 400 feet in width and formerly contained thick course gravel under overburden of 10 to

15 feet thickness. Values in the gravel were about 0.6 carats per cubic yard, not especially high but
consistent and large yardages were present. In the adjacent low terraces, values were somewhat lower,
but the overburden was shallower. The flat is now mined out, together with most of the payable
sections of the low terrace. The high terrace (23) has been only sporadically mined, and pitting by the
Geological Survey in 1961 showed that only unacceptably low values are present.

The gravels of the Gbobora flats were remarkable for their heavy mineral concentrates. Those
with good diamond values carried up to 200cc of coarse concentrate per cubic yard, the bulk of this
concentrate being black, purple and grey water worn corundum. Topaz and tourmaline were common.
Ilmenite, purple garnet, chrysoberyl, chromite, cassiterite and gold were also noted. About 300 cc of
fine concentrate per cubic yard of gravel was usual, and this concentrate consisted principally of
ilmenite and zircon, with some topaz, red garnet, chromite, rutile and gold, and rare grains of
magnesian ilmenite.

Values in the Gbobora appear to cut off just to the north-east of Njaiama, but reappear again in
two upstream tributaries, the Konta (25) and Kenyeyima (26) swamps. These swamps contain patchy
diamond concentrations with values up to 0.5 carats per cubic yard, and abundant corundum
concentrates. Near the head of the Kenyeyima, the gravels contain large quantities of very coarse
corundum, some of it well-rounded. In both swamps, most of the corundum is apparently of exotic
origin, but its distribution bears little relationship to the distribution of diamonds, and it seems fair to
assume therefore that the diamonds are of very local origin. Several grains of magnesian ilmenite were
detected in the fine concentrates from both swamps. In 1961 the Geological survey spent some time
endeavouring without success to find the course bodies. It was concluded that the sources were
probably small discontinuous kimberlite bodies below the swamps, but as the water table in the
swamps was high and they Survey at this time possessed no pumps, systematic trenching to check this
conclusion could not be undertaken.

The Masaye stream (27) flows south from Njaiama in a deeply incised valley whose floor is 200
feet lower than that of the Gbobora. Low, patchy values rearely exceeding 0.2 carats per cubic yard
have been found at a few mining sites between Boaia and Gondama in thick course gravels containing
much rock debris. Corundum with varying degrees of angularity is usually abundant but values
generally appear to be associated with the present of a few well-rounded quartz pebbles in the gravel.
Above Boaia, neither pebbles nor diamonds occur.

In the course of their regional sampling programme, the Diamond Exploration company found
kimberlitic heavy minerals in numerous streams in this block, and in the Masaye these minerals were
traced to their source. According to Stracke (1963), the course was a narrow kimberlite dyke about 20
feet in length, striking at 90° and cutting the bedrock of the Masaye flats about one mile upstream from
Boaia. No diamonds were found, either in the kimberlite or in the associated alluvial gravels.

Intermittent licensed mining has taken place along the flats of the Fayei stream (28) for about
one mile upstream from the Lease boundary. Many licence holders have failed to find diamonds in
workable quantities, and most of the others have only been able to recover between 0.1 and 0.25 carats
per cubic yard. Small well-rounded quartz pebbles occur in the gravel, and their presence is
inconsistent with the size of the present stream, which is 4th order. Kimberlitic indicator minerals were
found by the Diamond Exploration Company.

Parcels of diamonds from this Block are readily identifiable. Coated stones are rare; among the
clear stones, many have a brownish tinge, and a distinctive pale green colouration is also common.
Good octahedral are rare, most tones having a very irregular shape.

All the facts available tend to show that the diamonds of block 3 come from sources which are
within the Block and are as yet undiscovered. It seems probable that these sources either constitute an
extension of the Koidu-Yengema dyke zones or are related in some way to the Gborboa-Masaye fault,
itself an extension of the Oyie-Shongo fault, one of the significant structures of Yengema Field.
The part played by ancient drainage routes is not yet understood, since these cannot be reconstructed
with any confidence. The rapid headward retreat of the Masaye source along the major fault has
apparently destroyed the valley of a form large Gbobora tributary.


Block 4 – NIMIYEMA

This is the first of the important Sewa Valley blocks and is the furthest upstream, being situated
immediately to the south of the S.L.S.T. Lease boundary. The most important deposits it includes are
those of the Sewa channel and terraces, and there are no deposits within the Block which cannot be
related either to terraces or to old courses of the Sewa. The Sewa here drops steeply from the
Thousand-foot Surface to the Coastal Plain Surface in a series of rapids and falls, the most impressive
of which is Big Kongo Falls, near Punduru, in the stretch of rapids and minor falls between Big Kongo
and Ngawama, the river drops 200 feet, and further upstream at Jopowahun rapids there is a drop of
about 40 feet. Between these major rapids and falls, the remaining reaches of the river consist of mile-
long sections of very shallow gradient separated by minor rapids. The course of the Sewa, in this
Block is exceptional in that it is not closely controlled by a major fault, but it exhibits instead several
bends which are related to minor structures. An earlier deep valley of the river lies to the west of, and
parallel to, the present valley, but now contains only a small stream, the Yombi.

An examination of the Yombi valley shows clearly that, by the time the Sewa River abandoned
it, the steep nick-point in the river profile between the Coastal Plain and the Thousand-foot Surfaces
had advanced to a point just north-west of Nyandehun. In contrast, in the present profile, the
corresponding nick-point (Big Kongo Ngawama) is still far to the south of Nyandehun. As a result, the
floor of the upper part of the Yombi valley is at a much lower level than that of the nearby present
Sewa valley, and at Nyandehun village it would be possible at only moderate cost to divert the Sewa
through a short tunnel into the Yombi valley. If this was done, eight miles of Sew channel containing
an estimated 425,000 carats of diamond would then be dry, apart from the discharge of minor
tributaries, which could be channelled into leats where necessary. Unfortunately this is scarcely a
practicable scheme at present, because the security problems presented by the deposits of rich, easily-
mined gravel thus exposed would be insuperable.

A diversion scheme may eventually be of interest if a demand for hydro-electric power arises
within reasonable distance of this area. The Sewa, diverted at Hyandehun by a dam 90 feet high, could
be made to drop a vertical distance of 300 feet into the adjacent Yombi valley. These figures are based
upon a rapid compass, tape and Abney level survey of the Nyandehun watershed.

Mining of the Sewa channel, both by diving and by damming, has continued from 1956 until the
present with little diminution of interest, and recoveries have in general been high. Most pools have
yielded values in excess of 2 carats per cubic yard. One pool in the centre of Jopowahun rapids (39) is
estimate originally to have contained 800 cubic yards of gravel with an average grade of 10 carats per
cubic yard before it was mined in 1961. In 1962, the Geological Survey took random samples, 3 cubic
yards in aggregate, from the tailings dump beside this pool and recovered 5 diamonds of 2.3 carats total
weight. Numerous pools remain to be mined, chiefly the deeper and larger ones. One of the most
interesting contains an estimate 96,000 cubic yards of gravel under thick sand over burden and 8
fathoms of water.

The Sewa has developed no flood-plain in this Block, but there is one long strip of low terrace

(34) between Nyandehun and Jopowahun. Mining of this deposit has been rather sporadic because of
the combination of deep over burden with exciting values. On the other hand, high terrace deposits,
which are very extensive, have been closely prospected by the diggers and all the payable sectors
mined. The outstanding deposit of this type was undoubtedly the Kpakama terrace (37), where there
were large areas of gravel-in-place and disturbed gravel. Recoveries between 1.0 and 3.0 carats per
cubic yard were made from most of this ground, which is now worked out.
Most of the swamps in block 4 intersect high terrace gravels (41, 42, 43, 44 and 45) and
recoveries of 0.3 and 0.4 carats per cubic yard have been general. The Kpakama swamp (46), beside
the rich Kpakama terrace, appears to be a comparatively recently abandoned channel of the Sewa,
possibility comtemporaneous with the low terrace deposits elsewhere, and it contained thick well-
rounded quartz gravels below deep overburden. ~average values were good, about 0.5 carats per cubic
yard, and some pits gave recoveries of three or four times this figure. The deposit is now mined out.
The Bawi swamp (47) also contained moderately good values in normal thin angular gravel under
shallow over burden. Here there are no indications of terrace material on the adjacent hill slopes, but
the swamp gravels contained rare rounded quartz pebbles and some water worn corundum.

The flat floor of the Yombi valley (38, 39, 40) is largely occupied by a flat terrace in which the
gravel is overlain by between 6 feet and 12 feet of semi-consolidated over burden. This terrace is cut
by the flats of the under fitted Yombi stream and by several tributary swamps, but there is usually very
little difference in ground elevation. In the lower sections of the valley (39, 40) terrace flat and swamp
have all been continuously mined during the past ten years and there is now very little workable gravel
left. The upper section of the valley (38) which is two miles in length, contains poor values and has
been only sporadically mined. All the Yombi gravels are coarse, containing a high proportion of only
semi-rounded quartz cobbles. Well-rounded cobbles and pebbles are present, but are not as common as
might be expected, in view of the valley’s history.

In general, the diamonds of this Block are not markedly different from those of other parts of the
Upper Sewa. The median size of stones recovered from the river channel is about 1.1 carat, and
usually between 30% and 40% of the stones are coated. Few parcels were seen from the river below
Big Kongo Falls, but in these parcels the clear diamonds always appeared to be of exceptionally good
colour and were largely free of inclusions. In the Lower Yombi deposits, the median size of stones
recovered is about 1.5 carats, which compensates for the somewhat low average grade of the gravel.

Heavy mineral concentrates tend to be abundant in the Sewa channel, up to 4000 c.c. per cubic
yard having been noted during the present survey, and they contain about 85% tourmaline and 10%
corundum. Other minerals noted were reddish-purple garnet, ilmenite, zircon, rutile, magnetite and
gold. The same mineral suite occurs in swamp and terrace gravels, but the tourmaline/corundum ratio
is reversed in all those deposits which are not very close to the river. In the Yombi gravels, tourmaline
is rare. Magnesian ilmenite can be found in small quantities in the fine concentrates from most
deposits, but it is rare in the Yombi, fairly common at Kpakama (46) and very common in the Sewa
channel gravels near Jopowahun (29, 30).

The source of the diamonds in this Block, as elsewhere in the Sewa valley, is still a matter for
conjecture. Here, the evidence is rather conflicting. On the one hand, all deposits are related to the
Sewa river, past or present, and the diamonds bear a general resemblance to those of the Yengema
Field. On the other hand, there do appear to be slight variations in diamond characteristics between
individual deposits, and magnesian ilmenite, a kimberlitic indicator material, occurs in unlikely
concentrations in some of the Sewa gravels.

Block 5 – JAGBWEMA

In this Block, and also in Block 6, the Sewa flows through the straight and narrow Sewa Gorge,
which has an average width of about one mile. The floor of the gorge is part of the Coastal Plain
Surface, and the sides of the gorge are formed by the dissected scarp of the Thousand-foot Surface.
Below the village of Nyawama, the river has achieved a very stable mean gradient of about 2.5 feet per
mile and there are only very minor widely-spaced rapids or breaks-of-slope. The river channel has
begun slowly to swing, forming low-amplitude curves which are to a large extent uncontrolled by
bedrock structures. Under these conditions some flood-plain deposits have accumulated inside the
curves, but because of the extremely slow rate of meander formation at this gradient, the process has
periodically been interrupted by slight re-grading of the channel. As a result of this, most parts of the
flood-plain deposits are now slightly elevated with respect to the channel floor, and are therefore
classified as low terrace.
The Lower Falima Valley, through which the Sewa must formerly have flowed when it occupied
the Yombi valley (Block 4), contains extensive high terrace deposits (50) which underlie the villages of
Bambarakalima and Laborye. Stain, sub-angular quartz gravels occur beneath 10 feet of over burden,
and diamond values vary from 0.1 to 0.5 carats per cubic yard. Only sparse water worn corundum is
present in the concentrates from the terrace gravels, and there appear to be no rounded quartz pebbles.
This is very odd, unless the deposit is attributed not to the old Sewa but to the Falima itself, which
might have reworked the old Sewa gravels, removing all the pebbles but leaving some of the diamonds
behind. Much of the payable gravel has now been mined from this deposit, and although substantial
yardages remain they are barely payable and are partly covered by the two villages. The falima flats
have been little exploited; over burden is thick, and values are likely to be low.

Apart from the Falima deposits, no other high terrace deposits occur n this Block, and diamonds
are found only in the channel, flats, and low terraces of the River Sewa. No swamp or tributary
deposits have been found. The Sewa channel is the major producer (51 to 56) and has been intensively
and continuously mined since the inception of th Alluvial Diamond Mining Scheme. Sediment
deposition has been less erratic than is usual in the Sewa, and areas of river bed without gravel are rare.
Most coffer-dam operations have exposed four to five feet of coarse pay gravel below five to eight feet
of sand and barren gravel. Values also are less erratic than usual, falling generally between 0.5 and 3.0
carats per cubic yard.

In the absence of islands and rock-bars in this part of the Sewa, coffee-dams can only be
constructed in semi-circular or semi-elliptical form, based on the river bank. Furthermore, their exact
locations are controlled by the distribution of suitable sound foundations in the form of bedrock on
which sediment is absent or thin. In several sections of the channel, therefore, the presence of thick
sand on the bed precludes the construction of coffee-dams, and those that are built usually fail to
enclose the deeper parts of the channel. In consequence, in spite of ten years of intensive mining, there
still remain substantial areas of river channel which have been mined only by divers.

At the extreme south end of the Block, below Dambara, where the channel swings against the
western wall of the gorge, there is a long deep pool which has never been mined because of the depth
of water. Nothing is known of the bottom sediments. The channel here may well be rock-floored,
scoured of all sediment, but if not, any gravel present should contain above-average values. This pool
merits the attention of any mining company which has some capital resources, but before any major
expenditure on equipment is undertaken, experienced divers should be employed to inspect the channel
floor and the occurrence of gravel should be confirmed by platform-mounted Banka Drill.

Flats and low terraces (57, 58) have been fairly intensively mined in the vicinity of the principal
riverside villages, but only sporadically mined elsewhere, and considerable reserves of gravel therefore
remain. Values in the exploited ground have ranged from 0.4 to 1.2 carats per cubic yard.

The diamonds from most deposits in this Block have a median size of about 1.0 carat and about
40% of them are coated stones. The only marked exception to this general description is the low
terrace deposit to the north of Jagbwea (57), where coated stones constitute 50% of the total recovery
and the median size is about 0.4 cart. Course concentrates are relatively abundant throughout and
consist predominantly of water worn tourmaline, with from 5% to 15% corundum and a trace of
ilmenorutile. The fine concentrates consist largely of ilmenite with some magnetite and rutile, and rare
fragments of red and purplish garnet. Only one – 1mm grain of magnesian ilmenite was found in the
concentrates of this Block.

All the diamond deposits here are in Sewa sediments and appear to represent a simple
downstream extension of those in Block 4. although average values are higher than in Block 4, this is
probably due entirely to the shallow gradient. There is no evidence for the existence of local sources.

Block 6 – BARMA

This Block has been the outstanding producer of alluvial diamonds for many years. Total
production to the end of 1965 is estimated to have been almost two million carats, and the biggest part
of this has come from the northern half of the Block, the Konta-Barma area.
Much of the topographical description of Block 5 applies equally to Block 6. The Sewa Forge
remains straight and its width remains constant at about one mile, but in the river, mean channel
gradient has increased slightly to about 3 feet per mile and curves are fewer. The profile is interrupted
by several minor sets of rapids, though many of which there is a narrow fault-controlled main channel.
At two of them, however, river width has trebled and braided channels appear at low water; these are
the rapids at Konta (60, 61) and Barma (66 and 67) which have been the focal points of the extensive
mining activity.

Mining of the Sewa channel, prosecuted for many years with tremendous fervour, is now
diminishing steadily because of the increasing difficulty of finding accessible un mined gravels.
Recoveries from channel gravels were normally between 1.0 and 5.0 carats per cubic yard, though
occasionally higher, and in the Konta and Barma areas it was in fact hard to find barren virgin gravel,
even where the deficiency of other heavy minerals indicated that little concentration had taken place.

The wide Konta and Barma rock bars were covered to a large extent by boulder beds, gravel
banks, and sandbanks underlain by gravel. All these were exposed and accessible to low water and
were therefore among the first deposits to be exploited. Although they are now regarded as worked
out, only sand and tailings remaining, yet a few miners manage nevertheless each year to make small
recoveries, and it is a matter for speculation whether all the diamonds they recover were left in the
tailings by the early miners or whether some have been contributed by the river at high water stage.

Substantial reserves of un mined gravel remain in the Sewa channel in this Block, for the same
reasons that applied in Block 5, but they form a smaller proportion of the original resources. This is
because extraction of gravel has been virtually complete in certain areas where the deposits have been
exposed at low water or where the presence of rock bars and islands has assisted the construction of

Flat and low terrace deposits are of lesser extent than those in the Jagwema Block, but the values
in them are appreciably higher and they have therefore attracted more attention frm the miners. No
deposit remains untouched although some have as yet only widely spaced pits. Many licensed mining
sites in these deposits were tested by the Geological survey (62, 63, 68, 70, 72, 73, 75, 76, 77) and most
gave recoveries of about 1.0 carat per cubic yard; none gave less than 0.5 carat per cubic yard. A few
high terrace deposits occur, the largest being that immediately south of Konta, but they have as yet
been little exploited and values appear to be generally low. Only four small swamp deposits have been
mined, near Baram.

Little difference was observed between parcels of diamonds from different parts of the Barma
block. All consisted of what might be described as a typical Upper Sewa selection, the median size
being about 1.2 carats, 30% to 40% of the stones being coated, and, among the clear stones, slightly
flattened or distorted octahedral predominating with layered faces, few flaws and good colour. There
have been frequent reports of the discovery of large stones here, but the only one which could be
confirmed was that of a 95 carat clear stone, found near Baram in 1961 and seem by the Area
Superintendent of the Mines Division.

Heavy mineral concentrates, which occur in quantities up to 2000 c.c. per cubic yard, consist
principally of tourmaline but also contain about 20% corundum and some ilmenite, ilmenorutle and
purple garnet. In most deposits, one or two grains of magnesian ilmenite can be found by thoroughly
searching the fine concentrates from several cubic yards of gravel.

Although the deposits of Block 6 are a continuation of those in Block 5, there appears to be no
alluvial explanation for the higher values per square yard, and it is difficult to accept that all the
diamonds are of exotic origin. Some could well have come from a local source, and the most probable
situation for this is considered to be to the north of Konta, in or near to the river channel.
Block 7 – BOAJIBU.

At the northern end of block 7, the Sewa Gorge widens out to become the Sewa Valley, which
has a width of four to five miles but is still flanked by the dissected scarp of the Thousand-foot Surface.
The river still flows on the Coastal Plain surface, but the mean channel gradient has increased to 5 feet
per mile with the result that the course is largely controlled by the principal basement lineaments and is
predominantly straight. Average values in the river channel deposits are appreciably lower than in the
upstream Blocks, but values in the terrace deposits are fairly comparable. Tributary stream and swamp
deposits, which are notably absent from Blocks 5 and 6, are common here on the eastern side of the
valley, although rare on the west.

The Sewa channel here contains several rock-bars and small islands, and in consequence a great
deal of coffee-damming has been possible, particularly in the sector to the west of Boajibu (83, 84, 88).
As there has also been intensive diving in this central sector, and there are few pools deep enough to be
inaccessible to the divers, there is little payable grave left. Immediately upstream, extending from the
double bend to the northern Block boundary, is the three-mile sector of channel which was the subject
of an experimental dredging operation by the Diamond Exploration Company from 1961 to 1965.
(Forristal, 1965).

Prior to 1962, mining activity in this northern sector was never particularly intensive, but this
was generally attributed to the deterrent effect of large areas of moderately deep water. In the course of
the experimental dredging operation, bulk samples were mined from thirty separate sites, and the
section dredge demonstrated its ability to dig gravel, clean bedrock and recover diamonds at all water
depths between 5 and 60 feet. The mining operation itself is more fully described in Part I, section D


The results of the dredging provided valuable data on the variations in thickness and grade of the
Channel gravels, and established that diamonds could be effectively recovered from the irregular Sewa
bed by suction dredge. Unfortunately, they also established that overall grades were extremely low in
this sector and that the value of diamonds recovered could not even cover the running costs of mining.

At Gendema Ferry, the long deep pool was found to be rock-floored and scoured free of
sediment. Above the ferry, the channel floor was still mainly of hard rock with very little sediment,
and the average grade of the small yardages of sand and gravel which could be found was 0.007 carats
per cubic yard. Below the ferry, there were extensive gravel deposits of irregular thickness overlain by
sand. 63,093 cubic yards of sand and gravel were mined and treated giving an average recovery of

0.047 carats per cubic yard. In the best cut, the grade was 0.139 carats. Using estimated sand; gravel
ratio of 3½:1, the grade of the basal travel would appear to be about 0.2 carats per cubic yard overall
and 0.62 carats per cubic yard in the best cut.

The accuracy of these results is unquestionable, but they are anomalous and some explanation
must be sought. Even low terrace deposits in the Boajibu area usually give better recoveries than those
listed above. In the stretch of river immediately downstream from the dredged sector, recoveries by the
Geological Survey from License=-holder’s gravel, obtained by both diving and damming, have ranged
from 0.5 to 2.0 carats per cubic yard. It would appear that the three-mile sector sampled by the dredge
was an especially unfavourable one. Above the Gendema Ferry, the scouring of the channel floor is an
obvious adverse factor. Below the ferry, where the channel gradient is very low and the bedrock
irregular, there has been sediment accumulation under favourable conditions, and there is no apparent
reason for the shortage of diamonds.

In the southern part of the Block, from the elbow (88) down to the Block boundary, the Sewa
channel is very straight and much of it is narrow. Mining by diving has been rather sporadic, and there
have been only a few coffee-dams. Values in gravel extracted have, however, generally been about 1.0
carat per cubic yard (94, 96), and the low incidence of mining activity is due mainly to the difficult
river conditions. Much of the channel is very deep, and cannot be tackled either by diving or by
damming, whilst many of the shallower parts are rock-floored, with little gravel. There are several
deep ppools, and gravel reserves are known to exist in some, but nothing is known of the bottom
conditions of the majority, and it is to be expected that many of them are scoured of all sediment. The
pool immediately below the Bohun rapids is considered to be the one with the best diamond potential.
Along all sectors of the river, the banks and low terraces have been mined since 1958, and very
few are left which have not been exploited. Reserves of diamondiferous gravel remain, but only in the
form of pillars and patches in erratically-mined ground. Values are generally attractive, between 0.4
and 1.0 carats per cubic yard, but the deep overburden, makes the extraction of pillars, after the initial
random pitting, a most arduous task. It has only been attempted at the Paninga deposit (82) where
values were higher than average, and even here about 30% of the diamonds have been left un mined
and un- mineable. It is in these and other low terrace deposits along the Sewa that the unsuitability of
the current licensed mining techniques is most in evidence.

Several high terrace deposits have been discovered and mined, but the most important was
undoubtedly the Gendema terrace (79) where there were two payable gravel beds in place under thin
over burden. The upper gravel bed consisted of a semi-laterised fine pebbly gravel whose thickness
varied from 1 foot to 4 feet and which carried values up to 0.4 carats per cubic yard. Below this was a

5 feet thick bed of mottled semi-laterised sandy clay which carried a negligible heavy mineral content
and no diamonds. The basal gravel underlay this false bedrock and varied in thickness from 6 inches to

3 feet. It consisted of large and small quartz pebbles, most of them sub-angular, in a matrix of semi-
laterised clay, and it carried values which varied from 0.3 to 0.8 carats per cubic yard. It lay on a very
irregular bedrock surface of decomposed granite cut by numerous quartz reefs.

The high terrace deposit between the Sewa (84) and the Kamada swamp (86) was mined in 1960
by the Lebafric Company. The diamonds were found in disturbed alluvial gravel and the underlying
lateritic gravel to depths of three to six feet below the surface. The ground mined consisted of those
blocks which contained values up to 0.6 carats per cubic yard. Mining ended when these blocks were
exhausted, as the remaining ground carried only 0.1 or less, which could not cover mining costs. The
‘pay streaks’ were found to be randomly oriented.

Several tributary deposits have been mined in the Boajibu block, but the bulk of the tributary
production has come from the Kamanda swamp (86) and the Legbiye flats and swamp (90).

The Kamanda swamp has become something of a legend because of the high values and the very
large numbers of miners that worked there during the years 1955 to 1957. It is estimated that the
swamp gravels originally contained average values of about 1.0 carats per cubic yard, but diamond
distribution would have been uneven, as is usual in swamps, and recoveries much higher than this must
have been made from some pits. Stones up to 8 carats in weight were reasonably common, and larger
stones appeared occasionally. The swamp has been worked over several times, and no vestige of the
original gravel now remains. A few diamonds are left, but they are dispersed through tailings,
overburden and top bedrock. In recent years, a high terrace gravel residue has been discovered and
mined on the slopes to the south of the swamp, and in this deposit, values of 0.25 carats per cubic yard
occur in the top 4 feet of soil and lateritic gravel, together with sparse water worn corundum and rare
stained pebbles.

The Legbiye flats (90) contained a normal flood-plain gravel under fairly shallow over-burden.
Average values were only a fraction of those a Kamanda, but gravel yardages were very much greater,
and consequently years of steady mining have resulted in a comparable production. Considerable
yardages of original gravel remain, but these are the portions rejected during each episode of successive
mining and therefore not only contain unacceptably low values but are also interspersed with
accumulations of tailings and dumped over burden. There is no detectable high terrace deposit near the

Following the Legbiye deposits upstream, the values in the main flats cut off just before the
stream passes under the motor road, but low values persist up the roadside tributary swamp. These
values diminish northwards and finally cut off entirely as this swamp passes under the road. To the
east of the road, and to the east-south-east of Boajibu, the headwater branches of the swamp cross an
interestingly level crypto-alluvial area of about one-fifth of a square mile. This area is underlain by
thick uniform sand, in which a gravel probe will penetrate to a depth of 15 feet without bottoming. It
was hoped to investigate this area with a Banka Drill, but unfortunately, the opportunity did not occur.
The sand looks like a fluvial deposit, and it is therefore possible that it may fill a buried old channel of
the Sewa. If it does, then it is likely also to be concealing a deposit of diamondiferous gravel.
To summarise the alluvial mining picture in Block 7, it can be said that licensed mining activity
has been diminishing steadily for many years and is likely to diminish still further owing to the near-
exhaustion of accessible payable ground. Most of the small remaining known reserves are in the Sewa
banks and low terraces. Substantial additional resources may possibly exist in some of the deeper
Sewa pools in the southern part of the Block, and also in the conjectural buried channel to the east-
south-east of Boajibu.

Heavy mineral concentrates are considerably less abundant in the gravels of this block than in
those of the upstream Blocks, and the proportion of fourmaline has decreased 70% of the course
concentrates consist of tourmaline and 25 of corundum, together with kyanite, metamict zircon and
chrysoberyl. Fine concentrates contain ilmenite, zircon, garnet and epidtoe. Forristal (1965) records
the occurrence of haematite, pyrite, topaz and staurolite in the dredge concentrates, in addition to the
above mentioned minerals. No kimberlitic indicator minerals were found in this Block by the
Geological Survey, but the Diamond Exploration Company reported the occurrence of a few small
grains in two of their samples. Little can be said at present concerning the origin of the diamonds in
this Block except that they all appear to have been transported to their present locations by the Sewa


The known deposits in this Block are of little economic consequence as the reserves are very
small and production has also been very small. However, it is necessary for them to be separately
classified and described, as they are completely unrelated to those of the adjacent block 7.

The Kobonde stream carries diamonds from its confluence with the Bofioye for a distance of
about one mile downstream, but the only workable deposits have been those beside the village of
Kobundala (101). Here, values up to 1.2 carats per cubic yard were found in an area about 800 feet by
200 feet which included not only a section of the Kobonde channel andflats, but also a small swamp
opposite the village. These deposits are now mined out. One or two diamonds were found upstream,
in the Bofioye, but not in sufficient numbers to encourage mining. Heavy mineral concentrates at
Kobundala consist chiefly of white corundum exhibiting various degrees of wear, but they do included
a few small grains of magnesian ilmenite and pyrope. Tourmaline is absent.

At Simabu, a few diamonds only have been found in the swamp beside the village. At
Jagbaihun, there has been patchy illicit mining in the Kundubayei swamp (102), and the Geological
Survey carried out some pitting. Low values were present in a very thin sandy gravel which contained
only sparse heavy mineral concentrates in the form of angular corundum and metamict zircon. A few
grains of magnesian ilmenite were detected in the fine concentrates.

These deposits lie upon the thousand-foot Surface and are completely separate from, and
unrelated to, those of the Sewa. The diamonds are distinctive. Only 20% of those seen in the course of
the survey were coated, and amohng the clear stones, a high proportion contained abundant inclusions
or had a brownish cloudy appearance. Each of the deposits is considered to indicate the presence of a
small kimberlite outcrop in the immediate vicinity. About one mile south of Jagbaihun, a perfectly
circular crater-like feature was found, 400 feet in diameter, and occupied at the time of the examination
by a shallow lake. A line of Empire Drill holes across the lake showed that a few feet of mud and clay
were everywhere underlain by horn blende rich granite, only slightly decomposed and fairly hard. The
crater is assumed to have been formed by meteorite impact.


The Sewa Valley continues southwards through this Block, but the topographical limited of the
valley are steadily diverging and coming less well-defined. The Block includes numerous deposits in
Sewa tributaries, but production from them has been insignificant compared with the river channel and
terrace production, the greater part of which has come from the southern half of the Block.
North of Baoma, the river channel is very straight, with an average width of only 360 feet, and in
consequence there are several sections where little sediment has accumulated. Few coffer-dams have
been constructed here, and although there has been a great deal of diving in the past, this is now
considerably diminished. Accessible payable gravel is now difficult to find, but some reserves do
remain in a number of small deep pools.

South of Baoma, river mining has been fairly intensive. (111, 113, 118). Most parts of this
sector have been mined by coffer-dams at some time or another, but two or three dams are still
constructed each dry season as gravel extraction in the early years was often incomplete. One of the
most persistently mined parts of the channel (113), about 3,500 feet north-east of Sembehun, was
mostly floored with 12 feet to 15 feet of coarse gravel, all carrying good values. Large samples of this
gravel, exposed inside a coffer-dam, were washed by the Geological Survey in 1962, and recoveries
varying from 0.5 to 10.0 carats per cubic yard were made. Virgin gravel remains only in a few of the
deep pools in the southern sector of the channel.

Discontinuous strips of flat and low terrace occur all along the river, and south of Kponima these
have an aggregate width of 1,500 feet. (112, 114). All of them have been partially mined, but in some
cases, in the flat near Baoma for example, exploitation has been discouraged by the great depth of
overburden and large areas remain almost intact. There are two known high terrace deposits (110, 115,
117) but they contain pool values and have been only sporadically mined.

Although there are several deposits in tributary streams and swamps (104 to 109) they have all
contained poor average values and mining has everywhere been very selective. These deposits occur
where the tributaries cut through the Second high Terrace zone, and their diamonds, which are
accompanied by sparse alluvial indications, appear to be derived from terrace residues, although such
residues cannot be detects on the surround slopes.

During 1961 and 1962, the Diamond Exploration Company carried out an evaluation of the
partially-mined Upper Papayei Swamp (105). They found that although prospect pits were scattered
over most of the swamp, only small areas had been intensively mined. The angular basal swamp gravel
was thin, and often absent altogether , its place being taken by a coarse sand. Diamonds were found in
virgin basal gravel and sand, in tailings, and in over burden. Barber (1963) estimated that, within the
area evaluated, 1,407 carats remained in 102,258 cubic yards of mixed gravel, overburden and tailings,
an overall average grade of 0.014 carats per cubic yard. The average grade of 182 cubic yards of virgin
gravel washed was 0.01 carats, but since the values in tailings appear to be similar, there can be no
doubt that the grade of the gravel removed by the original miners was far higher than this. The
adjacent hill slopes were also sampled, but no diamonds were recovered from the 115 cubic yards of
lateritic gravel which were washed.

In the river channel and terraces, heavy mineral concentrates tend to be abundant, quantities
ranging up to 500 c.c. per cubic yard of gravel, and to consist principally of tourmaline, corundum and
ilmenite, other minerals being rare. The tourmaline/corundum radio is about 2:1. In the swamps,
concentrates are generally sparse, usually between 5 and 10 cc per cubic yard of gravel, and they
consist of water worn and angular corundum, angular ilmenite, and rare water worn tourmaline. No
kimberlitic indicator minerals were detected in any of the Geological Survey samples.

The diamonds of Block 9 differ slightly from those further north in that coated stones constitute
only about 25% of the total recoveries. Along the river, the median diamond size is about 1.0 carat, but
in the swamps, the median size is only 0.5 carat. All appear to have been transported by the Sewa
river, and there is nothing at present to indicate the presence of local sources.
Block 10 – LEVUMA

Here the Sewa River, hitherto flowing consistently southwards, changes its direction to west-
south-west. The Sewa Valley opens out and loses its identity as it merges with the main expanse of the
Coastal Plain Surface. The Block includes, in addition to a short sector of the Sewa, an important field
which is associated with two Sewa tributaries and extends from the Sewa to the Kambui Hills, 7 miles
to the east. It has been recognised for sometime that the diamonds in this tributary field are genetically
unrelated to those of the Sewa Valley, and the problem of their origin has attracted a great deal of
attention but remains unsolved.

The Sewa channel here has the unusually steep mean gradient of 10 feet per mile over the three-
mile stretch from the Matemu confluence to Naiama. Consequently, gravel deposition has been very
irregular, and in some parts there is none. Many years of diving activity and extensive dam
construction have resulted in the extraction of the gravel from all the more accessible pools, with
recoveries varying from 0.25 to 2.5 carats per cubic yard. A few deep pools remain intact, and gravel
may be present in some, but total remaining reserves in the channel are considered to be small. In
1963, Minerals Research Syndicate took Mining Licences over a number of deep pools on the western
side of the channel with the object of extracting the bottom gravels by means of two air-lift gravel
pumps. Unfortunately most of the pools were found to be scoured free of sediment, and the limited
deposits that were found consisted of non-enriched transit gravels containing only very poor values.
The enterprise was therefore wound up in July 1964, having failed to find any payable deposits.

Some small strips of flat are present beside the Sewa channel, but these have as yet attracted
little attention from the licensed miners. The extensive low terrace deposit between Naiama and
Yendema has only been sporadically mined because for many years the attention of the miners in this
area was concentrated on the high terrace deposit. (126). This contained both disturbed gravel and
gravel-in-place under shallow overburden and although overall values were only moderate, there were
the occasional small rich spots to stimulate interest. ~At one particular high terrace mining site, 1,000
feel to the north of Naiama, a total gravel thickness of 25 feet was recorded. Most of this was fine,
sandy, well-bedded gravel with little heavy mineral concentrate and no diamonds. At the base,
however, was an 18 inch bed of fairly coarse quartz gravel containing 0.25 carats per cubic yard, and
this was underlain by a very coarse angular gravel, mainly schist fragments in a clayey matrix, which
was only 6 inches thick and carried 0.8 carats per cubic yard. The diamonds recovered from this
deposit were of abnormally pool quality, 50% of them being coated stones and most of the clear stones
containing many inclusions.

The important field which extends eastwards from this part of the Sewa is generally known as
the Matemu Valley Field, but it can conveniently be divided into two sections for purposes of
description; the Gbatiye stream and tributaries, and the Matemu river and tributaries. Description of
the deposits in each of these drainage systems is followed by a discussion of the possible sources of the
diamonds in both.

The Matemu and tributaries
The Matemu is a 7th order stream which is diamond-bearing from its confluence with the Sewa
and Diema, where it emerges from the Kambui hills. (147). Although the valley continues back into
the hills, floored for another mile by flood-plain and high terraces, no diamonds have been found in
these valley deposits. Diamonds do occur in the hills, however, in a tributary of the Matemu at
Kagbwema, which is 2½ miles to the east of Diema and 250 feet higher up.

Information on the Matemu channel and flats is rather limited. Most sectors of the river have
achieved a stable shallow gradient and are forming flood-plain, consequently rock-bars are few and the
channel gravels are everywhere covered by sandy overburden. Under these conditions, neither diving
nor the construction of coffer-dams is a practicable mining method. The channel gravels are in fact
continuous with those below the flood-plain and are only likely to be extracted as part of a unified
mining operation for the whole deposit.
The flood-plain or flat is almost continuous along the river, being interrupted by only three
breaks-of-slope between Diema and the Sewa, and it has an average width of 600 feet. Gravel
thickness is variable, from 3 feet to nil, and the overburden, which in most sections consists of about 16
feet of running sand and silt, has so far prevented exploitation. Gravel samples could be obtained from
only three sites (150, 156, 158), and the average recovery from them was 0.6 carats per cubic yard. It
is unlikely that the licensed miners will ever be able efficiently to tackle the Matemu flats; if the
reserves have been correctly estimated, they would appear to be most suitable for exploitation by

Low terrace deposits do not occur along the Matemu. High terrace deposits of all types, on the
other hand, are very extensive but generally of low grade. Widespread sampling of the high terraces by
the Geological Survey in 1959 produced an average recovery of less than 0.1 carat per cubic yard, but
the pitting drew the attention of the licensed miners to these deposits. The miners began selectively to
mine certain areas, and their confirmed recoveries ranged from 2.0 carats per cubic yard at Koyama
(159) to 1.0 carat at Feiko (141), 0.5 carat at Devoye (154), and 0.3 carat at Kakayeima (153).
Continued exploitation has now resulted in the removal of virtually all high terrace ground with values
in excess of 0.1 carat per cubic yard.

The numerous deposits in Matemu tributary streams and swamps all contained good values and
were nearly all mined out some years ago. This was one of the first illicit mining districts; mining
began in the Feiko swamp in 1951 and spread rapidly with the result that, by the time licences were
introduced in 1956, a large part of the swamp reserves had already gone. All the swamp and stream
deposits intersect high terrace ground, and their gravels contain rounded quartz pebbles and abundant
water worn corundum.

The Boviye (142) is a 4th order tributary stream of the Matemu. The high terraces and rounded
quartz gravels which occur along it cannot be attributed to a stream of this order, and field examination
leaves no doubt that the Boviye occupies the former valley of a river of comparable size with the
Matemu. The Gbatiye deposits (128 to 137), which will be described later, represent an even earlier
course of the same river. Since the diamonds of the Matemu itself cut off at Diema, the present
distribution of diamonds can clearly be seen to correspond with the routes of this ancient river, rather
than with that of the present Matemu. It therefore becomes of great interest to follow these routes back
into the Kambui Hills.

A small diamond deposit occurs in a swamp tat Mufiahun (143), and about one mile to the
north-west of this there is a wind-gap. On the other side of the wind-gap is the Neaboye river. A long,
very obvious wind-gap joins the Neaboye valley to the valley of the Upper Matemu, at a pont where
there is a typical elbow of capture. These wind-gapes appear to mark the course of the ancient river,
and it seems therefore to have drained the same basins as the present Matemu-Neaboye system, the
principal difference from the present rivers being the point of emergence from the Kambui Hills.

In the centre of the long wind-gap between the Neaboye and the Matemu, mentioned in the
previous paragraph, lie the isolated diamond deposits of Kagbwema swamp (144). This swamp is sited
on the fluvial deposits of comparatively fresh appearance, whose present confirms that there is a wind-
gap here. Well-rounded quartz gravels, two to three feet in thickness, were found beneath 9 feet to 18
feet of grey silt and clay overburden. Average recoveries from the bottom gravels were high, about 1.2
carats per cubic yard, and stones of up to 20 carats were occasionally found. There was intensive
mining activity until 1960-, but the deposits are now worked out.

On the northern section of the wind-gap, between Kagbwema and the Matemu elbow, there are
more rounded quartz gravels but neither the geological survey nor the licences miners have been able to
find diamonds in them. A few diamonds are reputed to have been found in a terrace and a swamp near
the elbow (145), but sampling failed to confirm this. A brief outburst of digging at this spot resulted
from the discovery of large corundum concentrations but interest soon subsided. The high alluvial
deposit of Kagbwema is one of the most enigmatic in the diamond fields, but the problem of its origin
is part of the problem of the whole Matemu Valley Field.
The Gbatiye and tributaries

The Gbastiye is a 4th order tributary stream of the Sewa and its coarse runs 2 miles south of, and
parallel to, that of the Matemu river. From the mining centre of Levuma to the Sewa, very good values
occurred in all the Gbatiye flood-plain gravels, but the tributary swamps only carried values within a
short distance of the main stream. Original values were approximately 8 carats per cubic yard at
Nongoba (129), 3 carsts per cubic yard at Levuma (132) and between 0.25 and 1.0 carat elsewhere.
Above Levuma, values continued up various headwater branch swamps and were associated with a
wide zone of high terrace gravel residue. One of the most interesting deposits was the western end of
the high terrace (134) between Kako and Levuma. Here, moderate values were found throughout the
thick surface lateritic gravels, accompanied by water worn corundum. Although recoveries as high as

1.2 carats per cubic yard were made in a few spots, the overall average was about 0.3. In some parts of
the deposit values persisted to a depth of 8 feet in lateritic gravel, with the result that large yardages
were available for mining within a comparatively small area. This particular deposit is now completely
stripped of gravel, but to the east of Kako much of the high terrace ground remains untouched, because
values in this direction are in general very poor. Apart from this terrace, the entire Gbatiye drainage is
now effectively worked out, all swamps and flats having been worked over several times between 1952
and 1964.

All the gravels of the Gbatiye drainage carry occasional rounded quartz pebbles and much water
worn corundum, which establish that the Gbatiye follows the route of ancient major drainage. A broad
wind-gap to the south-east of Mendama indicates the route of the upper part of this old drainage, which
is thus clearly seen to be an alternative course of the ancient Matemu river. The Gbatiye terraces
contain only gravel residues and are at a higher level than the Boviye terraces (141, 142, 148), which
have gravel-in-place under overburden. The Gbatiye must therefore occupy the earliest valley of the
Matemu, subsequently diverted north-wards in the upper Boviye area (1432), where there is an elbow-
of-capture in the reconstructed old course.

The source of the Matemu Valley diamonds

Much work has been carried out on this problem both by the Geological Survey and by the
Diamond Exploration Company, but no conclusive answer has been produced. The diamonds of the
Matemu Valley Field are virtually all clear stones; coated stones occur but are rare. The median size is
about 0.8 carat. The dominant crystal form is the flattened octahedron, often glassy, sometimes with
inclusions. The rarity of coated stones effectively disposes of any suggestion that the deposits could be
related to an ancient circuitous course of the Sewa, none of whose deposits contain less than 205 coated
stones. Heavy mineral concentrates consist principally of water worn corundum and normal angular
ilmenite, with some zircon and occasional grains of rutile, chrysoberyl, ilmenorutile, columbite and
chromite. Tourmaline is universally absent.

The possible solutions to the problem of the source of the Matemu diamonds, together with the
objections to them, can be discussed under four headings:-

(i) Local Sources.

It may be that the apparent association with the ancient drainage routes is fortuitous and that the
deposits originate in fact from local kimberlite zones. Deposits not near the old drainage routes could
have remained undiscovered because of the deficienty of corundum in such localities. The chief
objection to this theory is that a very exhaustive heavy mineral survey of this entire field was carried
out by the diamond Exploration company, and no kimberlitic minerals whasoever were found.
Furthermore, inspection of the bedrock of many swamp, flat and terrace deposits by the Geological
Survey has disclosed no recognisable kimberlite. Large bodies of brown and purple clay somewhat
similar to decomposed kimberlite were found below the Gbatiye flats at Nongoba, and it was thought
that these might represent source-rocks, possibly meta-kimberlites. However, bulk sampling by the
Geological Survey proved them to be barren.
The absence of kimberlitic indicator minerals is not considered to be a serious objection, since it
is contended in this Bulletin that source-rocks which are deficient in indicator minerals exist in many
parts of the diamond fields, and the Malema deposits (block 31) can be cited as proof that such sources
do exist.

(ii) Unique source at Kagbwema

The course of all the Matemu diamonds could be at Kagbwema (144), which was strategically
situated near the confluence of the ancient Matemu and Neaboye rivers. Its situation, and the good
values found there, make this an attractive hypothesis. Once again, however, exhaustive sampling has
failed to detect a single grain of kimberlitic mineral. The Geological Survey investigated the bedrock
by Empire drill, and the results showed that a large part of brown clay about 250 feet long and 150 feet
wide lay beneath the deepest part of the alluvium in Kagbwema swamp. In view of the negative results
from somewhat similar material in the Gbatiye at Nongoba and of the difficulty of tackling the job with
the limited equipment at our disposal, no attempt was made to expose and sample this body.

(iii) Source in the Kambui Hills

It was once thought that the source might be in some part of the Kambui Hills, drained by the
Matemu or the Neaboye. This hypothesis seemed originally very feasible, so it was considered that the
source should present a relatively easy target for systematic exploration. The Geological Survey
carried out intensive prospecting of the whole Matemu drainage were found. Subsequently, the
Diamond Exploration Company undertook a thorough heavy mineral survey, not only of the Matemu
and Neaboye basins but also of all adjacent drainages. No trace of kimberlitic indicator minerals was
found. The joint results seem effectively to eliminate this hypothesis.

(iv) Source at Tongo

Most geologists who have studied this subject now believe that, however improbable it may
seem, the Matemu diamonds must have come from the nearest proved source area, which is Tongo
Field. This has always been the contention of Sierra Leone Selection Trust prospecting engineers, who
discovered the Matemu deposits in the course of their early reconnaissance work. In his final report on
the kimberlite prospecting operations of the Diamond Exploration Company, Stracke (1963) also
supports the hypothesis of Tongo origin.

The present Tongo kimberlite outcrops are at an elevation of 750 feet and are separated from the
Matemu Valley deposits by 20 barren miles of country and the barrier of the Kambui Hills, which have
peaks at 1,600 feet and passes (which might be wind gaps) at 1,200 feet. As the hypothesis is usually
stated, it is postulated that, on Tertiary planation surfaces at or above the level of the present Kabui
crests, the Kambui barrier did not exist and drainage from the Tongo area was to the south-west,
carrying large quantities of diamonds towards the Matemu and Lower Sewa areas. Following capture
of the Tongo drainage by a southerly-flowing stream, the diamonds in the Kambui area were no longer
being augmented from the source, but continued to be transported downstream in the beheaded south-
westerly drainage. In the course of the destruction of successive planation surfaces and the emergence
of the steep topography of the Kambui Hills, all diamonds are assumed to have bene flushed down the
developing Matemu drainage onto the Coastal Plain Surface, coming to rest in the Lower Matemu and
Lower Sewa.

In a recent refinement of the hypothesis, Haggard (1965) suggests that large accumulations of
alluvial and eluvial diamonds had formed on the late Tertiary surface in the Tongo area and were swept
off south-westwards in “gravel trains”, which formed in an arid climate and moved rapidly along
drainage lines as a result in the crustal tilt which he believes to have taken place at the end of the
Tertiary. The accelerated down-cutting of the Sewa following the tilt would have ensured that steep
gradients towards the Sewa were maintained for some considerable time. In connection with this
argument, reference should be made to Part II, Section B of this Bulletin, where the question of tilting
is discussed. There is no evidence that any tilting has taken place since the early Ecocene.
In support of the hypothesis of Tongo origin, it may be noted that the Matemu Valley diamonds
bear a general resemblance to the Tongo diamonds, as both groups include very few coated stones. A
direct comparison has never been possible, however, as the Matemu diamonds have only been
examined on the mining sites. There is no other supporting evidence, unless the absence of kimberlitic
indicator minerals in the Matemu area is considered as such.

By 1960, the Geological Survey carried out prospecting operations throughout the northern
Kambui Hills, covering all the areas over which the hypothetical drainage from Tongo might possibly
have passed. Neither diamonds nor any indication of ancient major drainage (e.g. rounded pebbles or
water worn corundum) were found. The absence of diamonds was confirmed by the widespread
occurrent of illicit prospect pits, none of which had been followed by attempted mining. It is not
possible to accept that a few million carats of diamonds could be carried across 20 miles of country by
any mechanism and leave no trace, and it is not, therefore, possible to accept that the Matemu
diamonds originated from the Tongo area. Derivation from local sources is the explanation preferred

Block 11 – YAMANDU

The Sewa has now completely emerged from the confines of the Sewa Valley and is flowing in a
shallow trough across the main expanse of the Coastal Plain Surface. The channel gradient is shallow,
about 3.5 feet per mile, but the channel has been incised to a depth of 100 feet below the surrounding
high terrace levels, and its alignment is still apparently rigidly controlled by basement structures. High
terraces are extensive, but there has been only insignificant formation of low terrace and flat.

Block II lies almost entirely within Baoma Chiefdom and it is, in terms of production and
resources, second only to Block 6. Production is less seasonal than in the Upper Sewa as there is, in
addition to a long streth of the Sewa, a wide zone of swamp and terrace deposits. The miners are thus
convenientlyl able to alternate river mining in the dry season with swamp and terrace mining in the

The river channel has numerous minor rock bars but the drop in elevation at each is very msall.
Intensive mining, both by diving and by coffee-dam construction, has been carried on for many and
still continues each dry season, although the number of divers is now severely diminished. Recoveries
are high, usually between 1.0 and 3.0 carats per cubic yard. There are several long deep pools which
nobody has tried to mine, most of them upstream from the railway bridge, nearly all of these pools
should contain thick gravel deposits. Substantial diamond reserves are therefore estimated to be
present, but a large part of them are in situations which render them inaccessible to licensed miners or
to the small, less well equipped mining companies.

The river flat is represented by the narrow bank deposits which are intermittently distributed
along both sides of the Sewa and are nowhere more than 100 feet wide, normally less. These have
been sporadically mined with variable results. Most of the excavations attempted never reach the basal
gravel because of the deep unstable overburden. When the gravel is extracted, values are usually
between 0.4 and 1.0 carats per cubic yard. Some narrow strips of low terrace have formed downstream
from the railway bridge but they have attracted little attention from the licensed miners except in one or
two localities (177, 178, 208) where there has been a little production from 0.6 carat ground.

High terrace deposits and swamp deposits are generally closely associated in this Block and it is
convenient to group the descriptions together. Three types of high terrace ground can be recognised:-

The terrace deposits here take the form of disturbed gravel or
gravel-in-place; they contain abundant well-rounded quartz pebbles and occur in wide belts alongside
the river. There are very large yardages with only low values, but there has been substantial production
from selective mining at certain localities (167, 168, 173, 175, 178, 190, 195, 196, 199, 207, 218) and
there is considerable scope for increased production by further selective mining of these deposits. They
are intersected by numerous swamps and small streams (168, 169, 170, 171, 174, 178, 189, 192, 195,
196, 206 in which the gravels ere uniformly payable, averaging 0.3 carats per cubic yard, but which
are now virtually exhausted.

(i) The First High Terrace Zone.
(ii) The Second High Terrace Zone. The only terrace deposits in this zone are gravel residues
dispersed in the soil over wide areas of subdued topography. The topography is dissected by
diamondiferous swamps (172, 176, 179, 180, 188, 197, 198, 200, 203, 204, 205, 209 to 213 and 219)
containing angular quartz gravels in which water worn pebbles are either rare or absent but in which
there is always some water worn corundum. Values in the swamps are extremely variable, from
practically nil to 1.0 carat per cubic yard (198) but in those which have been mined the average is about

0.2 carats per cubic yard. A little mining has been attempted in the lateritic gravel along the margins of
some of these swamps, but recoveries are always disappointing and tend to diminish upslope. The
Second High Terrace Zone clearly represents an older terrace than the First Zone, which is
characterised by recognisable alluvial gravel, but there is little topographical difference between them.
Both zones consist of low flat-topped hills and ridges separated by swamps, and although there must be
some difference in mean elevation, it is not apparent in the field.

This zone forms a broad belt to the south of Yamandu.
On superficial examination it would appear to correspond with the First high Terrace Zone, but it is
distinguished from that by some anomalous features for which no satisfactory explanation has yet been
found but which must indicate different conditions of formation. The terrace deposits form the crests
and slopes of low ridges which are separated by normal streams and swamps (182, 183, 184). They
occupy a broad east-west valley which is separated from the Sewa to the south by a block of higher
dissected country, but is connected to the Sewa in the east and the south-west by two areas of similar
low subdued topography from which terrace deposits are absent. The zone thus has the appearance of
an old curving valley of the Sewa.

(iii) The Ndogbogoma Terrace Zone.

The terrace deposits themselves consist of a sandy clay overburden, up to 5 feet in thickness,
which overlies a thin gravel of rather sparse angular quartz fragments in a lateritic clay matrix. Below
this is semi-lateritised decomposed granite bedrock, and the transition from gravel to bedrock is always
gradual rather than sharp, an unusual feature to find in a gravel-in-place deposit. The gravel in fact has
more the appearance of a residual rather than an alluvial formation. Another odd feature is the
complete absence of rounded quartz pebbles, which are normally considered to be diagnostic of terrace
gravels. Yet the gravel contains in addition to diamonds, up to 20cc of heavy mineral concentrates per
cubic yard, and these consist principally of water worn corundum with subordinate water worn
tourmaline. Both value and water worn concentrates continue down into the decomposed bedrock, and
the mining width may vary from 1 foot to 4 feet.

The Ndogbogoma terraces have been mined steadily since 1960, with an average yield of 0.25
carats per cubic yard, but they cover an area of nearly one square mile and almost three-quarters of the
original gravel yardage still remains. Much of this is in the form of rejected blocks in partially-mined
ground, consequently average values are now low or marginal, but there is still considerable scope for
selective mining, and production is expected to continue from these deposits for many years.

The terrace ground is cut by a network of swamps and streams in which high values were
general. Illicit mining began in these swamps in 1952, and thousands of miners were soon attractged to
Yamandu. Mining continued for several years and over 300,000 carats of diamond were produced.
Values ranged from 0.5 to 3.0 carats per cubic yard and were highest in the swamps adjacent to
Ndogbogoma village and in the flats of the Telo stream (182). All these swamps and streams are now
regarded as worked out, although an estimated 81,000 carats remain dispersed in tailings and
overburden. Sampling of tailings by the Geological Survey showed that the values in them were often
as high as 0.2 carats per cubic yard, which is sufficient to justify reworking.

Heavy mineral concentrates in the swamps and flats of this zone are generally fairly abundant,
the maximum of 150 cc per cubic yard being recorded from the gravel of the Telo flats, beside
Ndogbogoma. The coarse concentrates consist of water worn corundum and tourmaline, with a little
chrysoberyl, rutile and ilmenourutile, whilst the fine concentrates contain only zircon of various
colours. A few rounded quartz pebbles occur in the Telo stream gravels, but nowhere else.
Certain stream-flat deposits to the south of the Sewa appear to be independent of associated
Sewa terraces and must therefore be separately described. Gravels of the Segyei flat (192, 193 and
194) are diamondiferous from the confluence with the Sewa to the confluence with the Moteye swamp,
near the railway. Diamonds continue up to the Moteye swamp 9191 and 189) almost to its source,
which is near to the Sewa. The Moteye cuts high terrace at Kenyema, and the Segyei forms a boundary
to the high terraces of the Masalu area for a short distance; between these areas there is no visible high
terrace yet throughout the Segyei/Moteye deposits the gravels consist principally of smoothly-rounded,
bleached quartz pebbles. This drainage is therefore believed to follow an actual abandoned channel of
the Sewa which post-dated the youngest of the high terraces.

Mining sites are scattered throughout the entire length of the Segyei/Moteye deposits, but
mining has only been intensive in the vicinities of Masalu and Kenyema; substantial yardages remain
in the other sectors. Overburden is usually about 10 feet thick and has inhibited exploitation so far.
Many blocks of ground carry 1.0 carat per cubic yard, and ground with less than 0.25 carat is
uncommon. The heavy mineral concentrates of the Segyei gravels are of particular interest because
they contain numerous grains of magnesian ilmenite in the -1mm fraction.

The Kokoye stream (209, 214, 215, 216 and 217) is similar in many respects to the Segyei. It
does not intersect any detectable terrace deposits, yet its gravels contain abundant well-rounded quartz
pebbles and it is therefore also believed to occupy an abandoned channel of the Sewa. The deposits are
normal stream-flat gravels in which values are uniformly good, and all the licenses tested by the
Geological Survey were working gravel with values between 0.5 and 1.5 carats per cubic yard.
Although this deposit has been mined for several years, and the best ground is now gone, yet
appreciable reserves of payable gravel remain. A few grains of magnesian ilmenite were found in
concentrates from Kikoye gravels and from nearby swamp gravels.

Looking at Block II as a whole, it is noticeable that the composition of coarse heavy mineral
concentrates varies with the distance from the river. In the Sewa channel gravels, tourmaline forms
50% to 60% of the concentrates, and corundum, 40% to 50%. In the high terraces and swamps which
are furthest from the river, the proportions are generally 30% tourmaline to 70% corundum. All fine
concentrates consist mainly of zircon, with little ilmenite and rutile. Kimberlitic indicator minerals
occur in small quantities in the fine concentrates from the Sewa channel and in many of the swamp and
stream gravels to the south of the Sewa. They are completely absent, however, from all deposits north
of the Sewa.

There is some variation in diamond characteristics. In the Sewa channel itself, about 80% of the
diamonds are clear and 20% are coated; the median size is about 1.2 carat and the maximum size, 40
carats. North of the Sewa, in the Ndogbogoma deposits, only 15% of the stones are coated, the median
size is 0.8 carat and large stones are uncommon, the largest reported being 6 carats.

South of the Sewa, in the Segyei and adjacent deposits, the median size is only 0.7 carat, 40% of
the stones are coated, many of the clear stones have inclusions, and stones larger than 3 carats are very
rare. In the Kokoye, on the other hand, the median size is; 1.0 carat and occasional stones up to 12
carats in size are found; coated stones form only 10% of the total, and very few of the clear stones have

Although diamonds have doubtless been contributed to this Block from both the Upper Sewa
and the Matemu Valley, a considerable proportion of the diamonds present are thought to be of local
origin. It is not otherwise possible to account for local variations in diamond characteristics and the
presence of kimberlitic indicator minerals.
Blocks 12 – JOMU

A pronounced steepening of the Sewa profile gradient is apparent in this Block 9 miles of the
channel have a gradient of 5 feet per mile, and there is a steep, rocky, flume-like section where the river
drops 50 feet in three miles. This extended nick-point is the fore-runner of the advancing Bullom
Surface. Here, and at many other points downstream, the Sewa is actively degrading its profile as it
descends from the Coastal Plain Surface to sea-level. A small corresponding nick-point can be
observed in the Bebeye River, a short distance above its confluence with the Sewa.

The three mile rock bound section of the Sewa is adjacent to the village of Jomu. In the dry
season, the river is confined to a very narrow central channel which appears to follow a major fault
with a strike of 70°.

Just south of Jomu, both the fault and the channel are offset 500 feet by a shear zone which has a
strike of 10°. The offsetting of the channel takes the form of a z-bend. For some distance above and
below this bend, the channel is virtually scoured of sediment, except for small pot-holes scattered over
the rock-surfaces. At the Z-bend, however, a very deep pool has formed (225) which is estimated
originally to have contained 12,000 cubic yards of gravel with an average diamond content of 8 carats
per cubic yard. This is known as “Bragg’s Pool” after the English diver, E.I. Bragg, who discovered it
in 1961. During the 1962 dry season, it was mined by the Jomu Mining Company using airlift Pumps,
and about one-third of the gravel was brought up and treated. Further mining has been attempted in
subsequent years by other Native Companies, but work has been handicapped by the fact that each
year, in the rains, the river partially refills the pool with non-enriched gravel. As far as can be
ascertained, the bottom gravels have not been broached.

Most other parts of the Sewa channel in this Block have been intensively mined by divers, and
many coffer-dams have been constructed. Much recent activity has been concentrated in the subsidiary
channel south-east of Gondama (228). Mining here has been both by coffer-dam and by airlift pump,
and very little payable gravel now remains. Reserves in the main channel are also low. Only vestigial
bank and low terrace deposits occur in this Block, and they have attracted little attention from the

To the north of the river are very extensive high terrace deposits in which semi-lateritised
rounded quartz gravels are overlain by up to 8 feet of overburden. Although recoveries from these
deposits have been good, and their mining presents no particular problem, they have been exploited
only to a very limited extent, and this is largely because their presence is often unsuspected. So far,
mining has been mostly confined to the margins of formerly productive swamps (237 to 242).
Something of a problem is presented by the northern arm of the high terrace area (246, 247, 248 and
250). It is difficult to decide whether this should be attributed to the Sewa or the Bebeye river.

An unusual deposit exists beside the village of Maina (248). Here there is a deep depression in
the terrace bedrock, approximately 40 feet deep and 500 feet in diameter. It was originally filled with
inter bedded quartz gravels and sands, and diamonds were present in the bottom 15 feet of gravel.
Values increased downwards, reaching 5 carats per cubic yard in the basal gravel, and it is estimated
that about 25,000 carats were originally present. The deposit was mined in 1963 and 1964 by the Jomu
Mining Company, but it is not known that proportion of the diamonds were recovered.

Numerous swamps cross the high terrace areas, and all contain re-concentrated terrace material
with good values, recoveries between 0.2 and 1.5 carats per cubic yard being usual; all are now
worked-out. Values up to 3.0 carats per cubic yard were met in the central section of the Papayei
swamp (243).

The channel of the Bebeye river carries diamonds from the Maina area to the confluence with
the Sewa, but very little mining has so far been attempted. There is no flood-plain, and only very small
strips of low terrace, which have been mined at one site (250). Several strips of high terrace were
found beside the Bebeye and most of these have been partially mined, although values were rather low.
A small swamp, the Pupugbelu (254) cuts through Bebeye low and high terrace at Tungie. Overall
values in the swamp gravel, which is now worked-out, were 2.0 carats per cubic yard.
The swamps to the west of the Bebeye (231, 232, 257, 258, 259) all contained moderate overall
values originally, but the payable ground is now worked-out. Heir gravels contained occasional
rounded quartz pebbles and water worn heavy minerals derived from Sewa terrace residues, and these
residues have been detected and mined in the slopes flanking some of the swamps (232, 257, 258).

Throughout this block, between 60% and 80% of the coarse heavy mineral concentrates consist
of water worn tourmaline, the remainder being water worn corundum with a little rutile. The fine
concentrates are mainly of zircon, with some ilmenite, rutile and epidote. The only kimberlitic mineral
found was a single grain of magnesian ilmenite in concentrates from the Bebeye channel near Tungie.
There is very little variation in diamond characteristics between different deposits. The ratio of clear to
coated stones is everywhere about 4:1. Stones up to 12 carats in size have been found in several
deposits, and larger stones are occasionally reported, the largest confirmed find being one of 25 carats.

There are few problems connected with the deposits of block 12. Most of the deposits are
clearly related to the Sewa River, and there is nothing to suggest the presence of local sources. The
only possible question concerns the Bebeye river deposits, which take a wide swing away from the
known Sewa terraces. The Bebeye diamonds may be derived from the undetected residues of an
eroded terrace, but if so, one would expect to find deposits in most of the tributary swamps, and this is
not the case. Alternatively, the Bebeye could be following an actual old channel of the Sewa. A third
possibility is that there may be a diamond source in the Bebeye channel.

Block 13 – SEMBEHUN.

This Block is one of the most interesting in the whole field. Thirty-two separate swamp, stream
and high terrace deposits are disposed along a well-defined zone which runs parallel to, and three miles
from, the Sewa river. It has hitherto been generally assumed that this zone marks an ancient Sewa
valley, but the present survey has failed to find a single item of evidence which would justify such an
assumption. The deposits fall naturally into two groups; those of the Kponima-Mamboma area and
those of the Sembehun-Bangoma area.

The Kponima-amboma group. (261 to 275)

Many of the swamps within this group contained values which were patchy and barely payable,
but those in the vicinity of Kponima (268, 272, 274, 275) all contained gravel which gave an average
yield of about 0.4 carats per cubic yard, and recoveries of up to 2 carats per cubic yard were made in a
few spots. The values were present in normal swamp gravels of angular corundum, water word
tourmaline, rutile, zircon, ilmenite, orange garnet and chromate. Quantities were always low, rarely
exceeding 5 c.c. of concentrate per cubic yard of gravel. Rare grains of magnesian ilmenite, all less
than 1mm. in size, were found in the concentrates from the following deposits: 265,269, 270, 272, 274
and 275.

Several high terrace deposits have been worked. In some of them (271, 272, 275), angular
quartz gravel is overlain by between 4 feet and 10 feet of brown clay over burden. Heavy mineral
concentrates correspond with those in the adjacent swamps. These terrace deposits appear to represent
the flood-plains of former 4th or 5th order streams, which no longer exist in the vicinity. Average values
were not particularly high, and much ground of this type remains un mined. In th eother high terrace
deposits, values were present in the surface lateritic gravels, from which quartz fragments were
virtually absent.

At Kamabai (269), values of up to 3 carats per cubic yard were present in surface lateritic gravel
to a depth of 12 feet. A little coarse water worn corundum was present in the lateritic gravel. Within a
circular area about 180 feet in diameter, the licensed miners completely removed the surface gravels,
which lay upon a hard layered residual formation of gibbsite, quartz and sericite. Values were greatest
in the centre of this area and decreased towards the margins.
The wall of the excavation represented the limit of payable values, although a few diamonds were
present in the surrounding ground. Trenching by the Geological survey showed that the hard residual
“bedrock” was only two or three feet thick and was under-lain by several lenticular bodies of purple
micaceous clay, dipping at 40° to the south. Relationships with the enclosing decomposed granite were
obscure. Builk sampling of these formations established that no diamonds were present in them, so the
deposit must be an alluvial residue, but there is no way of relating it to any present drainage.

At the unnamed deposit (270) which was the scene of a phenomenal diamond rush in 1965,
diamonds were present in surface lateritic gravels and decomposed bedrock. The deposit was
discovered in mid-July 1965, and by early August the number of people on the site each day was about
4,000, although some of these were not miners but followed ancillary professions. The rich ground
was swiftly exhausted, and numbers diminished steadily after August, until by the end of the eyar only
about 200 men were still working. By this time, most of the surface gravel had been stripped from an
area measuring 1500 feet by 500 feet, to depths varying from 3 feet to 6 feet. Although some small
blocks of ground carried values up to 8 carats per cubic yard, there were only a few of these, and larg
yardages were in fact barely payable. Average values were estimated at 0.35 carats per cubic yard.

The area mined at site 270 is in the form of a broad belt which straddles a ridge between two
worked-out swamps. The crest of the ridge followd a dolerite dyke which has a strike of 120° and an
apparent thickness of 25 feet. The diamonds on the flanks of the ridge were present in surface lateritic
gravel and in the underlying semi-lateritised decomposed bedrock, a purple-red compact granular
material whose original nature is not known. Mining stopped at a depth of five or six feet, where this
red material had become too hard to mine and wash. On the crest of the ridge, the dolerite is well-
jointed; weathering has penetrated all the joints, which are now laterite-filled, and the texture of the
laterite indicates that it has formed by the replacement of dolerite rather than of introduced material. In
many places, some of this laterite has been soft enough to mine to depths of 8 feet below the surface,
and has been found to carry payable values.

Heavy mineral concentrates from the various diamond matrices at Site 270 have, almost entirely,
a grain size of less than 1mm and consist of ilmenite and zircon with a little magnetite and retile. Rare
grains of schomite and of doubtful magnesian ilmenite were noted. A trace of water worn corundum
was present in some samples of lateritic gravel, but only to the extent of one or two grains per cubic
yard. Rounded quartz pebbles are absent, and quartz of any sort is extremely rare in the +2mm wash.

Many questions are posed by this deposit and the Geological Survey has had no opportunity to
carry out the work necessary to answer them. It would be of interest to establish whether values persist
into the hard red bedrock formation and to investigate the exact mode of occurrence of values in the
dolerite joints. The presence of water warn corundum shows that fluvial deposits were at some time
present on an earlier surface, but the minuteness of the quantities of this corunudum make it
questionable whether the diamond concentrations can be attributed to any alluvial process.

In 1965, a kimbelrite exploration programme was begun by the Geological Survey, along and in
the vicinity of the Noniyei swamp (272). The programme was not completed, but it had met with no
success up to the time it was terminated by the author’s withdrawal from field work prior to retirement
from the Government service. The results are discussed in Part IV, Section D.

The Sembehun-Bangoma group (276 to 287)

This group contained, in addition to a number of small swamp deposits with poor values and
limited reserves, two long stream flats which became major producers because the large yardages and
the good values present. These were the Sembehun and Kakitoyei deposits. No high terrace deposits
have been found.
The flats and low terraces of the Bakwe (279, 280), a 5th order stream, carry diamonds for a
distance of 2½ miles and became famous as the ~Sembehun 17 deposit, after the village near which the
highest values occurred. The diamonds were present in the flood-plain and low terrace gravels, which
were angular, with rare rounded pebbles, and were under about 8 feet of over burden. Heavy mineral
concentrates consisted of water worn corundum and tourmaline, rutile, ilmenorutile, zircon, ilmenite,
chromite and actinolite. Mining began in 1965, and very high recoveries were made by the early
miners from one or two pits. The deposits are now regarded as worked out, but, from the results of
Geological Survey sampling in 1959 and 1960, average recoveries from the ground mined are
estimated to have been 1.2 carats per cubic yard, much of the production having come from erratically
distributed rich gravel pockets in bedrock depressions. On the other hand, an evaluation of the
remaining resources by the diamond Exploration Company in 1961 and 1962 showed that, within the
block of ground valued, an estimated 65% of the virgin gravel remained un mined and contained an
average of 0.4 carats per cubic yard (Barber, 1963). These contrasting estimates are not incompatible
in view of the fact that the bulk of the original reserves were contained in small bedrock depressions,
and bearing in mind the miners’ skill at selective mining.

The deposits in the flats of the Kakitoyei (282, 283) a 4th order tributary of the Bakwe, were
somewhat similar, although the stream is smaller and there is very little low terrace. Average values
were slightly higher and more evenly distributed than in the Bakwe. High values (0.5 to 4.0 carats per
cubic yard) were present in the lower 5000 foot section of the flats (283), whilst lower values (0.1 to

0.5 carats per cubic yard) occurred in the flats adjacent to Nyandeyama. The pay gravel consisted of
angular quartz with no rounded pebbles. Coarse concentrates were mostly of corundum, of which 20%
was water worn and the remainder angular, together with some ilmenite and occasional grains of water
worn tourmaline. Fine concentrates contained zircon, ilmenite, rutile, chromite, actinolite, tremolite,
and rare angular orange garnet fragments.

Within Block 13 as a whole, median diamond size is about 0.8 carat, and stones up to 15 carats
have been found. The proportion of coated stones varies from one deposit to another, from a maximum
of 30% coated (274) to a minimum of 5% coated (283). Throughout the Block, good glassy octahedral
are uncommon among the clear stones, which mostly consist of crinkled or distorted octahedral.

The author has formed the opinion that none of the diamonds of this Block have been carried by
the river Sewa. The deposits are found in a zone which begins near the Sewa and runs parallel to it, but
there is no evidence whatsoever to suggest that the Sewa ever passed this way. There is no noticeable
difference between the topography of the zone and that of the barren country to the north and south, all
of which are part of the Coastal Plain Surface. Rounded gravels are absent, except for a few pebbles in
the Sembehun 17 deposits, and these pebbles could well have been rounded by the 5th order Bakwe
stream itself. The few high terrace deposits and the widespread occurrence of water worn corundum
show that there has been much local redistribution of diamonds by small rivers which have followed
various routes during the development of the Coastal Plain Surface. Water worn tourmaline also
occurs frequently but cannot be adduced as evidence of Sewa influence, although it is abundant in the
Sewa gravels, since samples collected in the course of various prospecting traverses, during the present
survey, show that tourmaline in both water worn and angular forms is widely distributed throughout
this entire region.

Although there is no evidence of transportation of the Sembehun diamonds by the Sewa, neither
is there as yet any concrete evidence of a local source, but the distinctive characteristics of the
diamonds are difficult to account for by any other explanation.

Block 14 – WUBUNGE

Although this Block covers a 10½ mile sector of the Sewa River, little interest has been shown
by the licensed miners and production has been small. This is partly due to the fact that there are very
few diamond deposits in tributary swamps, but principally to the fact that average values in the Sewa
channel gravels are far lower here than in other reaches of the river.
The channel gradient here is 10 feet per mile. There are no major falls or rapids, but instead, a
continuous series of minor rapids and rock bars along the entire sector. In a typical dry season, 60% of
the area between the river banks consists of exposed rock. Gravel accommulations and therefore
discontinuous and erratically distributed, chiefly in small deep pools and potholes.

As in all other parts of the Sewa, the gravels from potholes and depressions on the exposed rock
bars have long ago been cleaned out. Some diving has been attempted in most parts of the channel, but
it has rarely persisted for long, as recoveries have never been particularly good and the divers have
soon migrated to more lucrative deposits upstream. A few small coffer-dams have been built during
the last seven years. There is no doubt that by far the greater part of the gravel reserves of this sector of
the Sewa channel remain intact, and although estimated values are low, they are adequate to support
selective mining, and these deposits may be expected to attract increasing interest as the other parts of
the river become worked out.

There are very few bank and low terrace deposits, and they have only been mined at two
locatilities (289, 292). There are several strips of high terrace, but these have been mined only in the
Folobu and Wubunge areas (293, 294, 295). At Folobu, the deposits consist of rounded quartz gravels
in place under 2 to 6 feet of over burden; at Wubunge, rounded quartz gravels containing soft laterite
nodules are covered by 10 feet of over burden. Four swamps intersect the high terrace at Wunbunge
and Folobu; these contained payable values in normal swamp basal gravel, but are now worked out.

No generalisations of value can be made about the diamonds of this block, because only an
unrepresentative few have been seen. They do no appear to differ noticeably from those of Block 12.
Heavy mineral concentrates consist of tourmaline and corundum, with ilmenite and zircon in the fines.
A few grains of magnesian ilmenite, under 1mm in size, were found in the concentrates from gravels in
the Wubunge area.

The diamond deposits of this Block would seem to be a simple alluvial extension of those in
Block 12, with the lower values appropriate to their downstream position.

Block 15 : PETEWOMA

This Block contains a group of swamp and stream deposits which have no particular alignment
or orientation. The area through which they are scattered is situated near the end of the 60° Sembehun
diamond zone of Block 13, and the deposits of the two Blocks are similar to that indications of Sewa
River influence are absent.

High values were found only in the vicinity of Petewoma town, in the headwater swamps of the
Kaputi stream (297, 298, 299, 300, 301), and in the Hewanjima swamp (296). In these deposits,
recoveries ranged from 0.25 and 2.0 carats per cubic yard from thin basal gravels of angular quartz;
they are now almost mined-out, but some reserves of ¼ carat ground remain along the lower Kaputi.

The flats of the Mogbui stream carry diamonds continuously for a distance of over 4 miles, but
over long sections, values are marginal or un payable. Mining has taken place at a space-out series of
sites, and recoveries between 0.2 and 0.8 carats per cubic yard have been made (303, 305 to 309).
Substantial reserves of gravel remain, but values in general are poor and it is to be expected that
selective mining will only produce another 5 or 6 thousand carats.

The small swamp at Kale (304) produced only a few diamonds, but it is of interest because it is
the only known diamond-bearing swamp which lies on rocks of the Rokell river Series.

Only one high terrace deposit is known in this Block (300), and it lies just north of Petewoma.
Values of 0.5 carats per cubic yard were present in an angular quartz gravel below 10 feet of
overburden. It represents the former flood-plain of a large stream, probably 4th order, which has now
disappeared. For high terrace deposits to be virtually absent, throughout an area which contains such a
large number of swamp and stream deposits, is distinctly unusual.
The possibility must be considered that they exist but have not been detected by the miners either
because of the pacity of corundum or because the diamonds are enclosed in laterite nodules.

Heavy mineral concentrates are not abundant in any of the Peterwoma deposits. A maximum of

40 cc per cubic yard was noted, but 10 cc per cubic yard was more usual in payable ground. Most
coarse concentrates contained both water worn and angular corundum, with variable percentages of
rutile and water worn tourmaline. In some cases, rutile made up the bulk of the concentrate. Fine
concentrates consisted chiefly of zircon, with subordinate ilmenite and rutile, and rare grains of a
brown garnet. No trace of kimberlitic minerals was found. Not a single rounded quartz pebble was

The number of diamonds examined here was not large enough to enable a useful general
description to be given, and no distinctive features were noted. The proportion of coated stones varied
from swamp to swamp, but the predominance of crinkled and distorted forms which was noted among
the clear stones of the adjacent Block 13 does not appear to apply in Block 15. Here, most of the clear
octahedral seen were well-formed with plane faces. There is nothing to connect the Petewoma
diamonds with the Sewa or any other deposit, and the presence of local sources must therefore be
considered probable.

Block 16 – HIMA

In this Block, the Sewa River crosses the outcrop of the soft Rokell River Series. The channel
form and the river deposits therefore differ in important respects from those which are found upstream.
The only deposits of the area are those of the Sewa channel and low terraces; there has been very little
mining attempted in these deposits and it is considered probable that there are very substantial reserves
of payable gravel.

Actual data are available only from mining sites in the low terrace so this will be discussed first.
Over burden is 10 to 15 feet thick and consists of brown clay. Below it, a bleached, sandy, rounded
quartz gravel lies on a bedrock of semi-decomposed friable shale. At the time of the examination,
gravel was only available from one licence at Mano Gbembeteye (311). Samples with an aggregate
volume of 5.3 cubic yards were washed by the Geological Survey and the mean recovery was 1.5 carats
per cubic yard, but examination of the tailings from other excavations indicated that this recovery was
higher than normal for this ground. Over 3 million square yards of un mined low terrace are present in
this Block, and in the absence of other data the gravel thickness of 18 inches, noted at Site 311, is
assumed to apply throughout. As a basis for the rough reserve estimation, the average gravel tenor has
been conservatively taken as 0.4 carats per cubic yard, giving a total diamond reserve of 600,000
carats. Some of the ground shown on the map as low terrace may in fact be flood-plain, in which case
it will have deeper over burden; in the absence of excavations it is difficult on the ground to distinguish
a boundary between the two.

The Sewa channel rarely departs from the graded condition as it crosses the easily eroded Rokell
formations. The present gradient is approximately 0.8 feet per mile, and it is somewhat difficult to say
why the river has not formed more extensive low terrace and flood-plain. Nothing is known of the
channel floor except that the river bed is of sand and is crossed by several dolerite dykes. There can be
little doubt that the sand s underlain by gravel, and it has been provisionally assumed that this also
carried values of 0.4 carats per cubic yard. Total diamond reserves in the channel are estimated at 1.2
million carats. It should be pointed out that the bedrock, consisting as it does of steeply-dipping shales
which strike across the channel, should have a surface whose minor irregulatrities must strongly have
favoured the formation of heavy mineral concentrations. Furthermore, accumulations of enriched
gravel may be expected behind the dolerite causeways which cross the channel. Needless to say, the
channel conditions preclude the use of divers and coffee-dams for mining, and there has therefore been
no exploitation of these deposits.

In the small parcels of diamonds examined, about 50% of the stones were coated, and most of
the clear stones were octahedral of good colour with very few inclusions. Stones of more than 2 carats
are said to be rare.
Heavy mineral concentrates in the low terrace amounted to about 100 c.c. per cubic yard of
gravel. 75% of the coarse concentrates consisted of tourmaline, 20% of corundum, and 5% of other
minerals – chrysoberyl, kyanite, ilmenite, chromite and orange garnet. Several grains of magnesian
ilmenite, all under 1mm in size were also noted. Derivation of these concentrates from the Rokell
Shales seems improbable, and it must therefore be assumed that they were brought by the Sewa from
Block 14. there is at present no reason to suggest that the diamonds have not also been so transported.
During any future mining operations, however, it is important that constant watch should be kept for
indications of kimberlite, since if it were to occur within the Rokell outcrop zone, it would be likely to
take the form of a major body.

It has been stated above that large diamond reserves are believed to exist in this Block. The
resources in the Sewa channel can only be recovered by a dredging operation, and for their exploitation
will perforce have to await the attention of one of the large mining companies; such companies are, of
course, excluded by present legislation. The low terraces are also suitable for dredging, but parts of
them can be tackled by the licensed miners. It cannot be too strongly emphasized, however, that it will
be a tragedy if the exploitation of the low terrace deposits is left of the licensed miners, since between
40% and 60% of the diamond reserves will then be permanently lost, in pot-holed un mineable ground.

Block 17 – SUMBUYA

This Block contains the last diamondiferous sector of the Sewa, together with numerous
scattered swamp and stream deposits. To the south of it, the Sewa flows sluggishly for another 20
miles down a well-graded sand-filled channel where over burden is very thick indeed and neither the
presence of gravel nor of diamonds has ever been established. The Sew River below Block 17 is not, at
present, regarded as part of the alluvial diamond fields.

The Sewa channel of Block 17 can be divided into two sections whose features contract. They
join at Sumbuya Ferry, where there is a change of gradient.

The 4½-mile section above Sumbuya Ferry has a fairly uniform gradient of about 8 feet per
mile, and contains a continuous sequence of rock-bars, islands and minor rapids. Channel width
oscillates between 1200 and 2400 feet, and braided low-water channels zig-zag between the banks;
these features result from the combination of steep gradient with absence of major structural controls.

Diamondiferous gravel occurs in large numbers of small pools, both deep and shallow, and as
deposits in pot-holes and depressions in rock-bars. Mining, by various methods, has been fairly
intensive. In recent years, diving activity has declined but pitting on the islands and the construction of
small coffer-dams still takes place each year. Recoveries have ranged from 0.4 to 3.0 carats per cubic
yard, but the higher values in this range are now rarely met. Considerable reserves remain in the
deeper pools, about one-third of which can easily be sealed off by coffee-dams, but overall values are
only expected to be about 0.5 carats per cubic yard, which is not very profitable for this type of

Below the Sumbuya Ferry is a 7 mile section of channel which is in a stable graded condition
with a uniform gradient of about 0.5 feet per mile; the river bed is of sand or fine gravel and is devoid
of rock-bars. The bedrock formations are gneisses of the Kasila Series, and the Kasila/granite contact
is believed to correspond approximately with the transition from steep to graded channel at Sumbuya.
No mining has ever been attempted in this section of the river, or in its banks or low terraces
consequently the presence of diamonds is not proved. However, a normal basal gravel is to be
expected below the channel sediments and in this situation, immediately downstream from the
confirmed deposits above Sumbuya, it would be anomalous if id did not contain diamonds. As a basis
for rough reserve estimation, an overall grade of 0.3 carats per cubic yard has been assumed for the
basal channel gravel, but it must be emphasized that this is a guess. An indication of possible actual
values could be obtained by sampling the deep flats which flank the channel two miles below
Sumbuya. Both these and the channel deposits are only suitable for mining by large dredge.
There has been no mining of Sewa low terrace of bank deposits anywhere in the Sumbuya block.
Values in the basal gravels of these are again provisionally assumed to be 0.3 carats per cubic yard, and
inspection of the river banks indicates that over burden is at least 25 feet thick. Sewa high terrace has
been mined at two sites. At Mofwe (320), recoveries of 0.35 carats per cubic yard were made from
semi-laterised rounded quartz gravels under shallow over burden. At the head of the Manjoma swamp
(321), recoveries of 0.25 carats per cubic yard were made from surface lateritic gravel containing
terrace residues.

Among the numerous swamp and stream deposits of Block 17, those to the west of the Sewa
appear to be directly related to river deposits, but those to the east do not. In the flats and terraces of
the Yanoi stream, recoveries varied from 0.2 carats per cubic yard near Mopende (315 to 5.0 carats per
cubic yard near the Sewa confluence (316), and the gravels contained a few rounded quartz pebbles.
Small reserves remain in the upper section of the stream, near Mopende. At Mofwe, the Gbegei flats
intersect Sewa high terrace, and near the village they were intensively mined for 8 years (319). Values
up to 2 carats per cubic yard were found in thick bleached, rounded quartz gravels below deep sandy
over burden. This section of the stream is now regbarded as worked-out, all the tailings have been
treated twice, but diamonds remain in some parts where the miners were prevented from reaching
bedrock by the instability of thick, sandy, water-bearing gravels. The lower section of the same plats,
towards the Sewa confluence, has received very little attention from the miners. Values are reported to
be poor but this may not be true, as the overburden here is very deep and unstable and the basal gravel
may never have been tested. The Gbegei flats at Mofwe are considered to occupy a former subsidiary
channel of the Sewa, corresponding roughly in age with the low terraces elsewhere.

To the east of the Sewa, no high terrace gravels whatever have been detected. No rounded
pebbles occur in the angular swamp and stream gravels, which originally contained erratically
distributed values ranging from 0.1 to 1.0 carat per cubic yard. Heavy mineral concentrates are
generally rather sparse; they consist of water worn and angular tourmaline, water worn and angular
corundum, zircon, rutile and ilmenite. Very high values are only reported from one small swamp (326)
near Yambama, but it was not possible to verify the reports as the swamp was completely mined out
and the tailings had been repeatedly washed. Concentrates here contained very little corundum or
tourmaline but abundant rutile.

Interesting features were observed in some of the deposits of the Malen stream. In the Kpatobu
mining area (329), the flats widen at the point where the stream intersects a 65° lineament which can be
send on the air photos. Values here were good, about 1.0 carat per cubic yard, and the median diamond
size was 1.7 carats, coated stones being rare. Upstream from this point only lowe values occur and the
median diamond size is about 0.5 carat. Downstream, values and diamond sizes decrease

Heavy mineral concentrates in this deposit are nowhere in excess of 10 c.c. per cubic yard, and
contain no water worn corundum. A small swamp, which branches off to the east from the main flat,
following the 65° line, also contained good values accompanied by only a trace of angular coarse
corundum. No kimberlitic indicator minerals were detected in the Malen deposits, but in view of the
distribution of diamonds, it is considered very probable that a diamond source of some type is
associated with the 65° lineament and is intersected by the Malen stream.

In Block 17 as a whole, the median size of the diamonds recovered is about 0.8 carat. Stones up
to 2.0 carats are reasonably common, and the maximum size in nearly every deposit is 8.0 carats,
although the recovery of one stone of 19½ carats from the Yambaye was confirmed by dealers’
receipts. The proportion of coated stones varies considerably from one deposit to another, but is on the
average about 20%.

The diamonds in the Sewa channel, terraces and western tributaries may well be of exotic origin,
with a long alluvial history, but those in the swamps to the east almost certainly are not. There is
nothing in any of these scattered swamp deposits to suggest that they are related in any way to the
Sewa or to any other major river. Two of the deposits described above (326, 329) are considered to be
particularly suitable for source rock investigations.
Block 18 – BAGBO

Five small swamp deposits containing low or marginal values occur in Bagbo, Chiefdom and
have been grouped together in Block 18. Average values in these deposits are less than 0.1 carat per
cubic yard, but selective mining has resulted in recoveries of 0.4 carat per cubic yard at Segbema (341),

0.3 carat at Mandu (343), and 0.2 carat elsewhere. The values are present in bleached angular quartz
gravels, but at Mandu these contain many small well-rounded pebbles, and at Segbema, one rounded
pebble was found after a search.

Heavy mineral concentrates in these deposits consist of columbite, angular corundum,
ilmenorutile, rutile, ilmenite and monazite, and are invariably present in only sparse quantities. A trace
of water worn corundum occurs at Mandu and Segbema, but within this Block, neither corundum nor
any other mineral is of use as an indicator mineral, and this is a formidable obstacle to successful
prospecting by the licensed miners. It is possible that several similar low-grade deposits exist in this
region, but most of them are likely to remain undiscovered.

Only a small number of Bagbo diamonds could be examined. The median size appears to be
about 0.7 carat, and the proportion of coated stones about 30%. No stones larger than 3.0 carats have
been found.

The Bagbo deposits are 10 miles from the Sewa river and there is no reason to connect them
with it. Although relatively flat, the area is now a watershed from which Sewa and Waanje tributaries
radiate, but the former presence of small rivers is suggested by the rounded quartz at Mandu and
Segbema. These may have effected some local redistribution of diamonds, but the absence of big river
influence is established by the extreme scarcity of water worn heavy minerals. It is assumed that the
Bagbo diamonds are derived from local sources, but no kimberlitic minerals could be detected.

Section E – TONGO FIELD.

Tongo field is a unit, a group of closely related deposits, yet for convenience in description some
subdivision must be attempted. The most important deposits are those of the Tongo river and its
tributaries. Of secondary importance are the downstream deposits in the Woa, and the peripheral
deposits of the Luya and the small Woa tributaries.


For most of its course, the Tongo River, a tributary of the Woa, flows in a shallow valley across
the 750 foot surface. A pronounced nick point occurs in the profile where it enters the Mining Lease in
the west, and this marks the scarp between the 100 foot surface and the 750 foot surface.

The flats and terraces of the Tongo carry diamonds over its entire length within the Mining
Lease, from the western boundary to the confluence with the Woa. Most of the tributary streams and
swamps also carry diamonds right up to their various sources, and the diamondiferous stream network
in fact drains an area of 20 square miles. This whole 6th order network flows on the 750 foot surface,
consequently profile gradients are universally low, and long sections of the stream are well graded and
have formed flood-plain deposits. The flood-plain gravels of the lower Lando and the Middle Tongo
all carry values in excess of 1 carat per cubic yard, whilst the gravels of the Upper and Lower Tongo,
the Kundo, the Allilooya, and Njaiye, the Gelia and the Weyakilo have occasional high-grade sections
interspersed with sections of lower-grade but nevertheless payable ground. Values in general are much
lower than those which obtained in Yengema field, and although small reserve blocks of 2,3,4 and 5
carats per cubic yard do occur, nothing comparable to the Woyie and Wongoyie deposits exists and the
average tenor of the known reserves in the Tongo drainage is little more than 1 carat per cubic yard.

The mining of parts of this drainage began in 1956. The deposits of the Lower Lango and
Allilooya streams are now worked out, but mining of the other tributaries and of the Tongo itself has
only recently begun, consequently the greater part of the gravels are still in reserve.
The Tongo drainage is cut by two kimberlite zones which both have a strike of approximately
60° true. The Sierra Leone Selection Trust Prospecting Department has traced the zones on the ground
by the use of heavy mineral sampling techniques, and trenches have been laid across the zones to
expose and sample the kimbelrites. The kimberlites are in the form of narrow discontinuous dykes
which resemble those at Koidu but contain higher values at outcrop. The Lando Zone, which appears
to be the most important, begins near Bumpe on the Upper Lando and follows the Kundo-Lando-Lower
Tongo line almost to Yendema, a length of almost 5 miles. This zone has been thoroughly explored
and several dykes have been traced, the average width at outcrop being 0.9 feet. It is almost certainly
the source of the diamonds in the Lando and the Lower Tongo.

One mile to the south-east of the Lando Zone lies the Middle Tongo Zone, which is only two
miles in length. Little work has so far been done on it and only two kimberlitic dykes have been
detected. The kimberlites of this zone may have provided all or part of the diamonds in the Middle
Tongo alluvials but little more can be said on this subject until the zone is known in greater detail.

No kimbelrite pipes have been found, and the heavy mineral sampling seems to show that no
other dyke zones crop out within the area of the Tongo drainage. This leaves several alluvial deposits
unaccounted for; those of the Allilooya, the Upper Tongo, the Njaiye, the Gelia and the Weyakilo.
The conventional view is that the diamonds in these deposits originiated from the known kimberlite
zones and were carried to their present sites by ancient drainage flowing southwards and south-
westwards. This explanation may well be correct in many cases, and it does receive considerable
support from evidence at Talama, in the Allilooya, and in the Luya that ancient river courses did in fact
follow some of the imagined routes. It must be pointed out however that there are, within Tongo Field,
a few isolated deposits, in some cases with high values as at the head of the Gelia swamp, which are
manifestly unrelated to any ancient alluvial diamond train yet are in areas where kimberlitic indicator
minerals are absent. The existence of such deposits convincingly suggests the presence of a source
rock which contains little or none of the normal indicator minerals. Sources of this type may well
account for many of these deposits which appear to be related to ancient drainage routes. If the
possibility is not kept in mind, important primary deposits could be overlooked.

(ii) THE WOA RIVER AND ITS TRIBUTARIES. (May Sheets 81 and 82)

The Tongo is a tributary of the Woa river, and several minor diamondiferous streams join the
Woa below the Tongo confluence. The Woa channel gravels are known to carry diamonds from the
confluence down, but values are believed to be low, and virtually nothing is known of the flats and low
terraces. Most of the diamonds in the Woa were probably contributed by the Tongo.

Small deposits occur in many of the minor Woa tributaries, but the four most important are the
Sandama deposits, the Boyei deposits, the Gandorhun deposits and the Classifaiya deposits. In the
Sandama area, payable values occur in several swamps and small streams, and they are related to a
kimbelrite dyke which, although apparently isolated, lies on the line of strike of the Middle Tongo
Zone, five miles to the south–west. Payable values are found in the Boyei stream and it tributary
swamps just to the west of Laoma. These might possibly be related to the Peyima kimberlite zone,
which is only one mile to the south.

In the Gandorhun area, diamonds occur in all the gravels of the Bwendai stream and its tributary,
the Peyima stream. This drainage is cut by a kimberlite zone which is three miles in length and from
which the diamonds of this area have obviously been shed. This is referred to as the Peyima Zone,
because trenching has shown that the dykes are best exposed near the Peyima, and a small pipe of
kimbelrite breccia, 20 feet by 50 feet, cuts the principal dyke near the point at which it crosses the
stream. Values are very low in the Lower Bwendai, but increase to payable levels in both branches
above the Peyima fork. Values in excess of 1 carat per cubic yard occur only in the Peyima flats near
the kimberlite zone, and the highest values of the deposit are found upstream from the kimberlite
Extensive low-grade deposits, with values not exceeding 0.1 carat per cubic yard, occur in the
Classifaiya stream and in some of its tributary swamps. These would appear to be related to the small
deposits on the opposite side of the QWoa and to the Naiagolehun deposits (381) outside the Lease
boundary, together with which they form a zone of approximate 60° strike. No kimberlite indications
have been reported.

(iii) THE LUYA DEPOSITS (Map Sheet 81)

The Luya is a third order stream at the south-west corner of the Mining Lease and a tributary of
the Mamaye. The gravels of the stream flat, which were mined out some years ago and yielded less
than 1 carat per cubic yard, contained numberous rounded quartz pebbles, showing that the stream
follows an old river channel. No name can be given to the river concerned since it was part of a
drainage system which has now vanished comopletely, but it is perhaps permissible to equate it with
the ancient river which can be shown to have flowed south from Talama (348) and to have been
subsequently captured and diverted through the present Allilooya valley. It is therefore conceivable
that this river may have brought the Luya diamonds from the Talama kimberlites, or even from the
Lando Zone. However, the existence of kimberlites in the Luya region is indicated by the occurrence
of kimberlitic ilmenite in the Siawur stream near Sewaoma.


Tongo Field comprises a large number of swamp, stream-flat and river flat alluvials, most of
which are grouped within an area of 40 square miles. Very high values are rarely met, but there are
large yardages of gravel with average values of about 1 carat per cubic yard. In the ten years since
exploitation began, approximately 1.6 million carats have been produced, and losses by illicit mining
have so far been tiny in comparison with those suffered in Yengema field; fortunately, some of the
most vulnerable outlying deposits were among the first to be mined by the Company.

The Company’s formal alluvial reserves on the 31st December 1965 were 1,512,100 carats.
Gross remaining alluvial diamond resources of the Field are estimated to be about 4½ million carats, a
figure which includes proved reserves, potential reserves and the substantial quantities of diamonds
which are dispersed in un payable ground. Production from alluvial deposits may be expected to
continue at current annual levels for 12 to 15 years, provided there is no increase in illicit mining.

As at Yengema, the long-term future of the field depends upon the possibility of successful
underground exploitation of kimberlite bodies. Whether the tenor and dimensions of those discovered
so far are adequate to support an underground operation has yet to be established.


Block 19 – BUNDOYE. (Map Sheets 80 and 81)

In spite of widespread prospecting activity during the past ten years there are still only four
known alluvial diamond deposits in this block. These trivial deposits would not have warranted
inclusion in a separate block were it not for the fact that they lie in an area which is of considerable
interest by reason of its situation between Tongo Field and the Middle Sewa.

It is widely known that the deposits of the Middle and Lower Sewa carry a higher proportion of
clear stones than those of the Upper Sewa, and many geologists believe that this is due to an influx of
diamonds from the Tongo area, where clear stones predominate. Stracke (1963) has suggested that the
earliest drainage from Tongo was east-south-eastwards and carried diamonds to the Sewa through the
area of Block 19.
Above Jalihun, the Maboa/Bundoye drainage flows on the 1000 foot surface. Gradients are low,
and most of the river channels are graded, with much flood-plain formation. Below Jalihun, the
Bundoye channel is rocky for a distance of 3½ miles as it drops 350 feet to the Coastal Plain surface,
which here forms the Sewa Valley floor. Diamonds occur at four points on this drainage. Above the
town of Dodo, the Maboa flats (345) contain low values of the order of 0.02 carats per cubic yard.
These appear to be a simple downstream extension of the Panguma alluvial deposits and attributable to
the present drainage. At Peyama (347), diamonds occur in a small swamp which intersects high terrace
gravels of bundoye, and which has been illicitly mined. Values appear to have been low. At Simbaru,
two ½ carat diamonds were found after protracted prospecting operations by local miners.

The most interesting deposit is that at Teyama (346). Diamonds occur in a small swamp which
occupies a windgap to the west of the Bundoye, and are associated with water worn corundum.
Sampling by the Geological Survey yielded one piece of grey-black bort of 3.2 carats. This deposit
appears to represent a re-concentrated terrace gravel of the Bundoye/Genduye, that is, that part of the
Bundoye system which flows from the north. To check the possibility of former eastward drainage
from the Panguma area, various swamps on possible wind gaps in the high watershed north of Dodo
were tested by the Geological Survey, and neither water worn heavy minerals nor diamonds were

As far as is known, there are no other diamond occurrences within this block. It is difficult to
accept that any major diamondiferous drainage could have crossed this area and left so few traces.
There is in fact no evidence for such drainage, diamondiferous or otherwise. Rounded quartz gravels at
Godi and along the Waye stream suggest that the primitive Maboa did at one time follow the obvious
south-westerly route down to the Matemu valley near Bado (Sheet 80, south-east corner). This
drainage may have had a tributary basin which drained both the Panguma and Tongo kimberlites, but it
is certain that diamonds from these sources could have been carried no further than Dodo, since along
the well-marked route from Dodo to Bado there are no diamonds whatever, either in recent alluvials or
in the rounded terrace gravels.

Whilst the early Maboa was flowing south-westwards to join the Matemu, the early Bundoye
was following a roughly parallel course which took it through the Teyama wind-gap (346) and along to
Peyama (347), which is on its present course also. The Peyama deposits are send to be related to the
Teyama deposits, which must themselves originate from a source either at Teyama or somewhere to the

Block 20 – PANGUMA.

In 1956, the town of Panguma experienced a diamond rush that has now become legendary, and
afterwards it remained for many years one of the busiest and best-known alluvial diamond mining
centres. The thousands of miners were attracted by the presence of large yardages of shallow, easily-
mined, rich gravel in one deposit; the Talama swamp (348). The Talama swamp is now worked-out
and mining activity has ceased, but in recent years the area has been the subject of continued attention
by the Geological Survey, as it contains dykes of diamondiferous kimberlite, the only proved examples
not covered by S.L.S.T. Mining Leases.

In the upper Talama swamp (348), shallow over burden was underlain by a thin gravel of
angular quartz, containing a few rounded pebbles and lying on bedrock of decomposed granite. The
average grade of this gravel is estimated, fro tailings sampling, to have been 5.0 carats per cubic yard.
The associated heavy mineral concentrates were very sparse and consisted of angular corundum, a little
water worn corundum, and some magnesian ilmenite, both coarse and fine. On the slopes tgo the north
of the swamp, lateritic gravel and disturbed high terrace gravel contained values which ranged from 0.1
to 2.0 carats per cubic yard. These slopes have now been stripped down to hard semi-decomposed
granite, except for a few scattered pillars of unwanted 0.1 carat gravel.

In the lower Talama swamp and other deposits to the south (349, 350 and 351), the deposits
were of a similar type to those of upper Talama, but values dropped progressively from north to south.
The Talama swamop finished at the Tongo stream which is barren upstream but continuously
diamondiferous downstream from this confluence.
A little to the north of upper Talama lie the Ngobiyei and Nyameina swamps (352). Values
Were high in the Nyameina swamp, which is now almost mined-out, but poor in the Ngobiyei, in which
fairly large yardages of barely payable gravel still remain. These two swamps meet and drain into the
Maboa stream, which is barren above this confluence.

Downstream from its confluence with the Nyameina/Ngobiyei, the Maboa carried values for a
distance of about 4½ miles. In the Panguma locality, the river, although graded, is incised a little
below the general topographical surface in a narrow steep-sided valley. Terraces are absent and the
width of the flat is erratic, with a maximum of 200 feet. The diamonds occur in the flat gravels, which
are rather coarse and contain a proportion of rounded gravel. Values are highest in the sector adjacent
to and just upstream from Panguma (353), where the original grade was 1.2 carats per cubic yard.
Further upstream, near the Nyameina confluence, values drop to 0.3 carats per cubic yard.
Downstream, values decrease progressively to 0.05 carats at site 354, and drop to completely un
payable levels as the Block boundary is approached. Some reserves of marginal ground remain in the
Maboa below Panguma, and mining continues there on a very limited scale.

Outcropping kimberlite dykes were discovered in 1960 by the Geological Survey on the slopes
above upper Talama swamp. During 1961 and 1962, a detailed programme of soil sampling, trenching
and kimberlite sampling was undertaken with the object of establishing the surface form, distribution
and grade of the kimberlite bodies. The final sampling work was somewhat handicapped by the
nocturnal activities of illicit miners, who endeavoured to strip off the weathered portions of the
kimberlite dykes immediately they were exposed by our trenching. In 1964 and 1965 the Division
carried out a programme of diamond drilling to determine the nature and dimensions of the kimberlite
bodies at depth.

The results of this work are fully described in Part IV, but a summary here will not be out of
place. The dykes, which are discontinuous, have an average thickness of 18 inches at outcrop and an
aggregate length of 4200 feet. There is no general increase of thickness with depth, and no major
bodies of kimberlite have been detected. Several dykes, which were intersected at depth, had not been
detected at the surface, but in no case did the thickness of these exceed 3 inches and they therefore
cannot be considered of economic interest. All the dykes dip steeply and are roughly parallel to one
another, forming a zone which has a strike of 65° true and which continues for a short distance into the
S.L.S.T. Lease area.

The Panguma kimberlite zone is clearly related to the nearby Lando, an middle Tongo zones, but
it does not appear to be an actual extension of either of these, although the line of strike of the Lando
zone, if produced to the south-west, passes close to the most northerly of the Panguma dykes. An
average recovery of 0.77 carats per cubic yard was made from bulk samples of weathered Pangum
kimbelrite, with a median diamond size of only 0.4 carat.

Median diamond sizes are higher than this in the alluvials, being 0.6 carat in the Talama swamp,
and 0.8 carat in the Nyameina swamp and the Maboa river. Moreover, diamonds between 1.0 and 2.0
carats in weight are reasonably common in the alluvials and all sizes up to 7 carats have been found,
but in the kimbelrites, stones larger than 1.0 carat are rare and the largest stone found weighed only 2.4
carats. The Panguma diamonds, both from kimberlites and alluvials, include only 3% by weight coated
stones and 4% bort. The remainder are clear octahedral forms of good colour, with occasional pale
yellow stones.

The total alluvial diamond resources of the Panguma area are estimated to have been originally
330,000 carats. The sampling results show that the known kimberlite bodies in the area together
contain about 180 carats per foot of depth and it therefore seems improbable, from a consideration of
quantities as well as the size ranges, that the Pangum kimberlites could have yielded all the diamonds
in the Panguma area. In fact, the local diamonds are believed to have been substantially augmented by
river-borne diamonds from the east.
Various items of evidence uncovered at Talma during the kimberlite investigations leave little
doubt that a river coming from the general direction of the Tongo kimberlites, flowed from east to west
across the Panguma kimberlite zone whilst planation of the 1000 foot Surface was taking place. These
items of evidence include patches of rounded gravel in various situatins, water worn corundum grains
on the eastern watershed, and a deep buried channel beside Byameina village, and they permit the
variations in the old drainage routes to be reconstructed. The earliest river flowed from the north-east,
and its drainage basin must have included the main Tongo kimberlite zone. After various changes of
course in the Talama area, discharging sometimes to the west and sometimes to the south, it was
eventually captured east of Talama and thereby permanently diverted away from the area, leaving large
concentrations of Tongo diamonds stranded around the Panguma kimberlites and along the present
Maboa valley. Later still, as the 1000 foot Surface of the Tongo area was stripped away, to be
superseded by the 750 foot surface, the entire drainage pattern was radically modified and all vestiges
of the old river disappeared. Around Panguma, however, the 1000 foot surface is preserved and with it
the diamond concentrations of mixed origin left by the vanished river. These have been reworked but
not significantly displaced by recent minor drainage.

Block 21 – FOINDU.

The deposits of this Block are in the flats and terraces of the Mamaye river and in three of its
tributary swamps; the most important in each category are those in the vicinity of Foindu town. Foindu
itself lies in a corner of the Coastal Plain Surface, and is overlooked by the dissected scarp of the
Tongo Surface. In this area the scarp is drained by the Mamaye, a 5th order stream.

The main product from this Block has come from the river terraces to the south-west of Foindu
and the swamps which intersect them. The extensive terraces consisted of high terrace gravel-in-place
overlain by shallow over burden (359), and a strip of low terrace with gravel under 12 feet of over
burden along the margin of the Mamaye flats (355). In both, thick semi-lateritised rounded quartz
gravels contained average values of about 0.22 carats per cubic yard, but recoveries up to 1.0 carat per
cubic yard have been attained by selective mining. In these deposits, large yardages of un-payable
gravel still survive in the form of pillars, but in many spots a bed of rich basal gravel also remains,
undetected below a false bedrock of semi-lateritised clay.

Very good values were found in the middle Mafaye swamp (356) and the lower Limaboye
swamp (360), both of which intersect the high terrace deposits. The swamp gravels were the first
deposits to be discovered and exploited in the Foindu area in 1957, and interest did not spread to the
terraces and the Memaye river until 1959. Values are poor in the upper Limaboye and there is very
little terrace beside the swamp, although the swamp gravels still contain rounded pebbles. The head of
the Limaboye, at the south-west corner of the Block is separated from the head of the beeya (Block 22)
by a level high terrace deposit in which the basal gravel has been only thinly and sporadically

The recent deposits of the Mamaye river can be considered in three distinct sections. In the first
section, ,which is that above Foindu, the channel is tortuous but nevertheless has a gradient of 50 feet
per mile as it cuts into the scarp of the Tongo surface. There is no flood-plain and, since the channel is
so steep and rocky, very little alluvial gravel of any sort. A few small patches of rounded terrace gravel
exist beside the channel. There has been sporadic mining in these and in some gravel pockets in the
channel but recoveries have been poor, never exceeding 0.3 carats per cubic yard.

In the second section, beside Foindu, the river has reached the Coastal Plaint Surface and the
Profile has abruptly levelled-off. The river meanders through a segment of flood-plain which is 1000
feet wide and 5000 feet long. Nothing is known of this deposit, except that probe-bar testing has
shown that the overburden, in which the water table is igh, is greater than 15 feet in thickness.
However, in view of the good values in related terrace and downstream deposits, it is a reasonable
assumption that the Foindu flats are underlain by payable gravels. As a basis for rough reserve
estimation, values of 0.4 carats per cubic yard and a gravel thickness of 15 inches have been assumed
to apply.
It is expected that these will eventually be shown to be very conservative. At the time of writing, no
attempt has yet been made to mine the Foindu flats, because attempted prospect pits have never
succeeded in reaching gravel. The deposit should in factg be of interest to a small mining company,
which would be able to undertake damming and diversion of the river into a leat and excavation by

The Mamaye flats downstream from the Foindu enlargement constitute the third section of the
Mamaye deposits. Here there is a gradient of 2.5 feet per mile and a flood-plain of varying width has
formed almost the entire section. The values are present in the flood-plain gravels, which consist
principally of sub-angular quartz with a few rounded pebbles, not but terrace deposits have been
detected. The over burden is from 8 to 10 feet in thickness, and miners equipped with pumps therefore
have little difficulty here. Licensed mining of these flats only began after the apparent exhaustion of
the terraces and swamps nearer Foindu, but now continues steadily every dry season with mrecoveries
between 0.4 and 0.5 carats per cubic yard. Values tend to diminish in a downstream direction, and no
diamonds have been reported below Yumbum.

The only other deposit of interest is the Nomiyama swamp (361), a Mamaye tributary about
1½miles to the north-west of Foindu. Here, values of about 0.3 carats per cubic yard were present in
swamp and high terrace gravels which both contained some water worn corundum but only angular
quartz. Although there is evidence here of an old drainage route, the nature of the gravels indicates that
the old drainage was only a 4th or 5th order stream, not a large river. This swamp does not represent an
laternative route of the river which laid down the Luya deposits one mile to the west.

The diamonds of Block 21 are nearly all clear octahedral, with a median size about 1.0 carat.
Coated stones are very rare; none were seen during the present survey, although it is reported that they
do occur. Heavy mineral concentrates consist chiefly of coarse corundum, both water worn and
angular, together with fine ilmenite, zircon and chromate.

The distribution of terrace deposits and rounded gravels shows that, above Foindu, the Mamaye
approximately follows the route of an ancient river, almost certainly the one which flowed through the
Luya valley. Initially, and apparently for a very long period of time, this river turned south-westwards
at Foindu and passed across the area now occupied by the Limaboye (360), through the wind-gap, and
down the valley of the present Beeya (Block 22). During this phase, it deposited the high terrace
gravels of the Foindu area.

River capture eventually took place just south of Foindu, and the river was diverted to the north-
east, into what is now the lower Mameye valley. This was the last change of course before the river
vanished entirely from the scene, and it resulted in the formation of the low terrace strip near Foindu
and in the dispersal of diamonds down the valley towards Yumbum. It is not clear what part the old
river played in the formation of the flats at Foindu. The present Mamaye appears to be underfitted in
these wide, deep flats, but this cannot be so, as it is inconceivable that there could have been absolutely
no lowering of base-level at Foindu since the old river vanished. We do not know how long ago this
was, but the time elapsed has been sufficient for many of the traces of the river to be removed by
erosion in the country to the north.

The distribution of diamonds in the Foindu area is manifestly related to the courses of the old
river, and this relationship suggests that the diamonds were brought into the area by the river. This
could be correct, but there is some difficulty in attributing the Foindu diamonds to either the Tongo or
the Panguma kimberlites in view of the wide intervening barren areas and the patent inadequacy of
these sources. In the opinion of the author, the large diamond deposits of the Foindu area can only be
attributed to a cource body in the immediate vicinity.

Blcok 22 – BEEYA

The narrow continuous Beeya/Kenja diamond deposit runs directly from the edge of the
Foinduarea to the Moa river on a constant bearing of 205°, the remarkably straight axis of the deposit
being offset only once, by the curve at Hangha. Block 22 contains the northern half of this deposit, in
which the diamonds follow the 205° line along the Beeya, a Kenja tributary, while the Kenja itself
diverges to the west and is barren.
For most of its six miles, the Beeya is a 3rd order stgream with very few tributaries, flowing with
a uniformly gentle gradient across the coastal Plain Surface and severely under-fitted in a wide flat. It
becomes apparent upon inspection that only a minor part of this flat is actually recent flood-plain of the
Beeya; the greater part of it is really the ancient flat of a former river and is therefore more accurately
described as low terrace. However, the two deposits can normally be considered as one along the
Beeya, as lateritisation of the older formation is not far advanced, the difference in elevation of the
bedrock surfaces is negligible, and there is no appreciable difference in average values. In the drier
years, the Beeya ceases to flow from February until May.

The Beeya deposits can be divided into three sections; north, central and south. The northern or
headwater section (362) is adjacent to the village of Kangama and consists largely of very wide
swamps which cover the low terrace deposits. The gravel, which is covered by 7 feet of clay over
burnden, contains very coarse but sparse water worn corundum and values in the tregion of 0.3 carats
per cubic yard, in a glutinous clay matrix.

In the central Beeya (363, 364, 365) the deposit varies in width from 500 to 1500 feet and
consists of unswampy low terrace and stream flat. The gravel contains a high proportion of sticky clay
and is covered by about 9 feet of over burden. Values are erratic; there are large blocks of ground
where the gravel carries between 0.1 and 0.3 carats per cubic yard interspersed with smaller blocks
where values up to 0.8 carats may occur. Because of the exceptionally tedious nature of the work
involved in mining these clay-rich gravels, a great deal of the lower-grade ground remains unmined.

In the southern or lower Beeya (366, 367) conditions are similar to those in the central Beeys
except that values are a little higher, and much of the gravel is clay-free and hense more easily treated.
During 1961 the Sierra Leone State Development Company (later to become theDiamond Exploration
Company) carried out an evaluation of the 3000 foot section of the Beeya deposit (366) which lies
immediately upstream from the main road crossing, near the village of Sembuhun 14. this was a well-
known mining area, regarded at that time as worked-out. The evaluation was carried out by means of a
series of deep trenches across the formerly mined ground. All the material excavated from the trenches
– tailing, over burden top bedrock, virgin gravel – was treated to recover the diamonds, and a great deal
of useful data was obtained. The whole project has been described in detail by Barber (1961).

The work showed that the alluvial miners had extracted only 60% of the diamondiferous gravel
bed. The average grade of the remaining gravel was 0.38 carats per cubic yard, and its thickness varied
from 6 inches to 32 inches, with an average of 12 inches. In his report, Barbar gave the following
reserve estimates for the swamp area of 350,000 square yards:-

Diamonds in virgin gravel remanants – 13,000 carats, valued at £140,000

Diamonds in mixed tailings + overburden – 18,000 carats, valued at £132,000

Hence, total remaining in swamp – 31,000 carats, valued at £272,000,

In approximately 1 million cubic
yards of mixed ground (=0.03cts/cu.yd)

Critical examination of the sampling procedures and results leaves no doub thtat these stimates
are essentially correct, and the deplorable wastage of diamond resources that sometimes occurs is thus
amploy demonstrated.

Total original diamond reserves in the area evaluated were estimated by the Geological Survey,
after sampling licensed miners’ gravel in 1959 and 1960, to have been approximately 70,000 carats.
Comparing this figure with the S.L.S.D.C. results suggests that the licensed miners extracted 60% of
the gravel, containing 80% of the gross diamond reserves. During their treatment of this gravel, they
recovered 70% of the diamonds present, giving a final net recovery of 56% of the total diamond
resources of the swamp.
Some small strips of high terrace gravel occur beside the lower Beeya, but values are mostly
very low indeed and there have been only sporadic attempts to mine them. One area near Ngelhun,
containing high terrace gravel residue, has been intensively mined (369) and recoveries averaged 0.25
carats per cubic yard. There are only three small tributary swamp deposits along the Beeya valley, and
these all appear to have been formed by the reconcentration of terrace gravel residues.

The diamonds of this Block are almost all clear octahedral, with a median size of 0.8 carat;
coated stones are rare. Stones up to 3 carats in size are reasonably common, and the maximum report
diamond weight is 10 carats. Heavy mineral concentrates consist of water worn and angular corundum,
chromite and zircon, and occur in quantities up to 200 c.c. per cubic yard of gravel.

There is abundant field evidence to show that the Beeya faithfully follows the course of an
ancient river, of at least 7th order, which can be readily traced back to the Foindu area and thence, via
the Mamaye, to areas further north. The occurrence of alluvial ldiamonds is restricted to the course of
this river, and this suggests that they were all carried into the area by fluvial transport. The ultimate
source was probably the same as that of the Foindu diamonds, and this is believed to be in the Foindu

Block 23 – KENJA

This Block contains the southern half of the narrow continuous deposit which extends from
Foindu to the Moa river, and it begins at the Beeya/Kenja confluence, which conveniently corresponds
with the 1:5,000 Sheet boundary. Above this confluence the Kenja gravels are barren; below it, they
contain diamonds all the way to the Moa.

The Kenja is a 5th order stream and is not so obviously underfitted in its valley as the Beeya, but
there are many terrace deposits of rounded gravel not attributable to a stream of this size, and these
show that the valley is that of a much larger river. With the exception of two tributary swamps,
diamonds have been found only in the flats and terraces of the Kenja valley, which differs again from
the Beeya valley in that the distinction between flat and low terrace of similar elevation can readily be
made in the field. Moreover, the flat valley floor is flanked by high terrace deposits of all types.

The gravels of the flat and low terrace all contain a proportion of well-founded quartz pebbles
and are generally free of clay. Grades vary; the highest values recorded during sampling were in the
flat gravels at Pava (380), which contained 1.5 carats per cubic yard; elsewhere, licence-holders
recoveries have been between 0.2 and 1.0 carats per cubic yard. There seems to be no completely
unpayable ground, as the poorest gravels yield 0.15 carats, which will give the licensed miners a bare
living. These deposits have been mined steadily for several years, but in spite of this, substantial
reserves remain and licensed mining can be expected to continue here for another six to eight years
with production decreasing annually.

The high terrace gravels were almost ignored for many years but began to attract the miners’
attention in 1960 and are now regularly mined at several sites during each wet season. The deposits are
of various types; semi-lateritised gravel-in-place under shallow over burden often exists on the lower
slopes near the flat, whilst gravel residues may cover the upper slopes or adjacent planed hilltops. At
Pava, there are several high terrace levels. Those on the middle and lower slopes and gravel-in-place
under shallow over burden and are now almost mined out; there is also an extensive upper terrace
surface whose lateritic gravel contains pebbly alluvial residues which have not yet been exploited. All
along the Kenja valley there are reserves of high terrace gravel, and although values are generally low,
selectric mining based on corundum concentrations enables the miners to make reasonable recoveries,
consequently mining is expected to continue for several years.

The heavy mineral concentrates of the Kenja are remarkable for their abundance. Quantities
between 100 and 500 cc per cubic yard of gravel are usual and on one rock bar (377) south-east of
Hangha, 2000 cc per cubic yard were present in the Geological Survey sample. These concentrates
consist principally of corundum, both water worn and angular, together with some chromite and zircon,
and traces of epidote, actinolite and topaz. No kimberlitic minerals whatsoever are present.
The Kenja diamonds are all clear stones, with a median size of about 0.7 carat. Generaly they
are of good quality, but it is noticeable that in all deposits in the vincinity of Pava a majority of the
stones are marred by inclusions. Stones of up to 15 carats have been found in the Kenja flats, and a 57
carat stone was recovered from high terrace ground beside Nyanyahun Bwema, which is on the eastern
side of the Kenja at the north end of the Block.

The Kenya deposits appear to be a simple downstream extension of the beeya deposits, the
diamonds having been brought from the north by the old river whose former course the Beey and Kenja
now follow. However, the occurrence of occasional large stones suggests that the exotic diamonds
may have been augmented from local sources. At Pava, a few small fragments of material lwhich
closely resembled weathered kimberlite were found in high terrace gravel tailings, but it was not
possible to locate the parent bedrock exposure in the overturned terrace ground.

Block 24 – LOWER WOA

This Block covers the minor deposits which exist around the south-eastern margins of Tongo
Field, and these deposits comprise only the channel, flats and terraces of the Woa river, and the
Naiagolehun swamp.

The Woa is an 8th order stream, and its valley represents a longue of the Coastal Plain Surface
which penetrates the dissected margins of the Tongo Surface. The channel gradient is only 3.5 feet per
mil,e and much flood-plain is now forming. There are few rock-bars or rapids and transport of coarse
and heavy material appears to have been at a minimum for some considerable time.

The Woa channel contains values, but these are not very high in the Wilima area and diminish to
negligible levels as the confluence with the Male is approached. This, and the difficulties of working
this sand-filled type of channel,, have discovouraged licensed mining activity. The few rock-bars have
been thoroughly cleaned-out of course, a few small coffee dams have been attempted, and there has
been sporadic diving activity, but the aggregate production has been small.

Flats and low terrace occur along most sections of the valley but have been mined only in the
Wilima area (382, 383). Here, thick gravels of rounded quartz lie beneath over burden which varies in
thickness from 10 feet to 25 feet, and they carry values that are occasionally as high as 0.6 carats per
cubic yard. These deposits are now regarded as worked-out, but virgin gravel remains in the deeper
parts of the flat and in pillars in the low terrace. Downstream there are large areas of untouched flat
and low terrace; nothing is known of the values, but for the purpose of resources estimation, an average
of 0.1 carats per cubic yard has been assumed.

The wide strip of high terrace near Wilima is now also mined-out. This contained rounded
gravels-in-place below shallow over burden, and included a few small high-grade patches among large
areas of barely payable ground. Below the south-east corner of the S.L.S.T. Lease boundary, values in
the high terrace gravels (384) are very poor, averaging 0.02 carats per cubic yard, and there is little
scope even for selective mining.

The diamonds of the Woa deposits are mostly clear, with a very small percentage of coated
stones, and a median size of about 0.6 carat. JHeavy mineral concentrates consist principally of water
worn corundum, together with chromite, zircon, and rare small grains of magnesian ilmenite. The Woa
diamonds are generally assumed to have been carried by the drainage from the kimberlite zones of
Tongo Field.

The only other deposit of importance in this Block is the Boyei swamp (381) near Naiagolehun.
Here, diamonds occurred n the swamp gravels, in which values averaged 0.2 carats per cubic yard over
a distance of 8000 feet. The swamp is now mined-out, but mining continues in the terrace gravels
along the sides and at the head. These terraces have angular quartz gravels below 7 feet of overburden
and represent the flood-plain of a former 4th order stream which must have played a principal part in the
distribution of diamonds both here and along the Classifaiya stream.
However, there is no trace here of a large river which might have carried the diamonds from a distant
source, and it is probable that the source of the Boyei diamonds is in the immediate vicinity. The few
diamonds seen were all clear stones. The heavy mineral concentrates contain ilmenourutile, sub-
angular corundum, ilmenite and zircon; no kimberlitic minerals are present.

Block 25 – — MALE

The deposits of Block 25 are those of the channel, flats, terraces and minor tributary swampos of
the Male river, from and to its confluence with the Moa.

The Male is a 9th order stream, and it flows across the Coastal Plain Surface with a mean
gradient of 4 feet per mile. In spite of the low gradient, there has been only insignificant formation of
flood-plain and terrace, and furthmore, the course of the river deviates only factionally from a
remarkably constant 204°. Both of these features must be attributed to the presence of a single
basement lineament, of regional importance, which exercises an overriding control over channel
alignment and inhibits lateral migration. The strike of this lineament is, of course, a common structural
direction in the diamond fields, one which is followed by many diamondiferous streams.

Information on the Male channel is available only at a few widely-spaced points. Diamonds
were present on all the rock-bars, which are few, but these have long ago been cleaned out, and no
guess can be made as to the former values. Diving and coffer-dam construction have only been
attempted at a limited number of separate sites in the southern half of the Block (389, 391, 393, 395).
The reason for this lack of interest must be the unsuitability of the channel for either method of mining,
since wherever gravel is extracted it always scontains good values. At the mining sites which were
examined, recoveries between 0.5 and 2.0 carats per cubic yard were being made, and similar values
were found in the gravel below the narrow strips of river bank. Considerable reserves of
diamondiferous gravel remain in the Male channel, the greater part of them covered by sandy
overburden in long pools of moderate depth. To a large extent, therefore, they are most suitable for
extraction by dredge.

A few strips of low terrace exist along the Male but very little mining has yet been attempted,
even though moderately good values are believed to be present. Very few high terrace deposits are
known, and mining has taken place at only two sites (393, 394), where values are low but still
comparable with those in many of the Sewa high terraces. It is possible that much more high terrace
gravel exists along the Male valley, and remains undetected because of the low level of mining and
prospecting activity in the region. Certain of the deposits in the minor tributary swamps (387, 392)
represent old channels of the river. Other swamps (394, 396) are cutting and reconcentrating high
terrace deposits. Recovereies between 0.2 and 0.4 carats per cubic yard have been made from these
swamps, but only unpayable ground remains now.

Heavy mineral concentrates in the Male deposits contain corundum, chromite, ilmenite, zircon
and orange garnet. Grains of magnesian ilmenite occur in the fine fraction of samples from the souther
section of the river (391, 393, 395).

There is some variation in diamond characteristics. At certain site (389, 390, 394, 395), from
10% and 15% of all parcels consist of coated stones, whilst at others (391, 392, 396) coated stones are
rare. At all sites, however, the quality of the clear stones appears to vary with the size. The smaller
stones, those under 1.5 carat, are almost all clear white octabhedra with few inclusions; the larger
stones, from 1.5 to 8.0 carats, frequently contain many inclusions and sometimes exhibit a brownish
tinge. The median size is about 0.8 carats.

The origin of the Male diamonds would at first glance appear to present no problems. There is a
continuous trail of diamonds leading from Tongo field, via the Woa and the Male, to the Moa river.
Diamonds are absent from the Male gravels above the Woa confluence, and very scarce indeed in the
Moa above the Male confluence. These are the unmistakeable features of an alluvial dispersion train,
and it was by following it that Tongo Field was originally discovered.
An inconvenient fact, however, is the proportion of coated stones in somme of the Male deposits,
which is very high in comparison with the proportion in the Tongo sources. It is the opinion of the
author that, although it almost certainly exists, the alluvial dispersion trail from Tongo to the Moa has
become very attenuated by the time it reaches the lower Male, and that many of the diamonds in this
river must be attributed to a less distant source.

Block 26 – PUTEHUN.

The Putehun Block contains three widely separated deposits which cannot easily be related
either to one another or to any other group of deposits. These are the Folu deposits, near Putehun, the
Kpaye deposits, near Sami, and the Yumbu deposits.

Near Putehun itself, the Folu flats (398, 399, 400) have been mined ssteadily for several years.
The Folu here is a small 3rd order stream which is manifectly underfitted in wide deep flats. On
aminisation, the flats are seen to represent the flood-plain of a former large stream of about the 5th
order. Over burden of 9 feet average thickness coveres a clay-rich angular quartz gravel, but the
occurrence of one or two well-rounded pebbles in this gravel suggests that the earlier stream either
followed or intersected the ancient course of an even large river.

Downstream from this deposit, in the Sami and Talia areas, some attempts have been made to
mine the Folu gravels but these attempts have always been short-lived and it is not known whether any
diamonds were actually found. Following the deposit upstream, it is found that values cut off abruptly
at a point about 4000 feet above the railway bridge; this cut off point is also the point of intersection of
a 65° lineament which is clearly visible on the airphotos and which is parallel to th etongo and Koidu
kimberlite zones. Heavy mineral concentrates consist principally of angular and water word corundum,
with subordinate columtie, ilmenite, zircon and chromite. No kimberlitic indicator minerals could be

About 15% of the Folu diamonds are coated stones; the median size is about 0.7 carat and the
maximum size, 8.0 carats. In view of the quantity of coated stones, the Folu diamonds can certainly
not be related either to the Tongo diamonds or to the nearby Kenja diamonds, but must have been
derived from an independent source; the source might well be associated with the 65° lineament
mentioned above. Earlier drainage seems to have been important in two ways; firstly, it produced a 2½
mile dispersal of diamonds southwards, and secondly, it furnished the concentrations of corundum
which stimulated prospecting activity by the diggers and caused the deposits to be discovered.

The deposits near Sami are found in the flats of the Kpaye stream (403, 404) and in one of its
headwater swamps (401, 402). The Kpaye is not underfitted, but the discovery, after intensive search,
of three rounded quartz pebbles in the stream gravel shows that there has at some time been a river
through this area. The swamp contained values between 0.2 and 0.5 carats per cubic yard, whilst the
stream flat gravels yielded from 0.3 to 1.2 carats per cubic yard. Bopth deposits are now worked out.
Heavy mineral concentrates consist of sub-angular conrundum, chromite, zircon and rutile; kimberlitic
minerals are absent. The Kpaye diamonds are mainly clear octahedral with very few inclusions and a
median size of 0.8 carats. Coated stones are rare, but about 8% of the stones have a pale yellow colour.

In the Yumbu stream (397), one diamond of 0.12 carats was recovered from 4 cubic yards of
flood-plain gravel during Geological Survey prospecting work in the Komende/Vaahun area. The
gravel contained abundant corundum, some of which was water worn, and numerous rounded quartz
pebbles. Prospecting upstream from the discovery site produced only negative results.

It seems probable that diamond deposits exist in many other swamps of this Block but remain
undiscovered because they happen not to contain corundum.
Block 27 – SEGBWEMA

This large Block contains 85 square miles of the Coastal Plain Surface east of the lower Male. It
contains a few scattered low-grade deposits and a 21 mile section of the Moa river.

This section of the Moa river lies immediately upstream from the Male confluence, and it flows
across the Coastal Plain Surface with a channel gradient of only 3.5 feet per mile. The Diamond
Exploration Company record the presence of kimberlitic indicator minerals in all the samples they took
from the Moa channel, but there is no mining activity there, and no diamonds were found in a 20 cubic
yard gravel sample taken from a rock-bar by the Geological Survey in 1961. It has been reliably
reported, however, that one diamond was found in one of several rock-bar samples during the
reconnaissance prospecting by Sierra Leone Selection trust.

In the swamps, diamond mining is known to have taken place at six widely separated places, but
there could be a few more sites, illicitly-mined, which this Survey did not discover. At Teyama (405),
recoveries of 0.3 carats per cubic yard were made from one small section of the swamp; in other
sections, values were negligible. The angular quartz gravel contained rare rounded pebbles and about
5cc of water worn corundum per cubic yard. The few diamonds seen were all clear octahedral of good
quality, with the unusually high median size of 1.5 carats. At Hanjama (406), mining has been very
sporadic and only a few diamonds have been found.

Dambu swamp (407) is an old mining site where nothing whatsoever is known of the original
values, except that they were certainly low. It was investigated by the Diamond Exploration Company
in 1963 after the discovery of one grain of chrome diopsite and two grains of magnesian ilmenite in
their heavy mineral samples. The swamp itself was trenched to determine whether kimberlite occurred
in the bedrock, but nothing except granite-gneiss was exposed. Loam samples were taken on a 200
foot grid covering the adjacent slopes and although no more kimberlitic indicator minerals were found,
one sample contained a 0.3 carat diamond. This might have been worth following up.

In the small swamp near Gandoruhun (408), values of 0.2 carats per cubic yard were present in
the angular quartz gravel, accompanied by 60cc of water worn corundum per cubic yard. All the
diamonds were clear stones, a few of them pale yellow in colour, and the median diamond size was
about 1.2 carats. The swamps near Sembema (409, 410) contained values between 0.1 and 0.4 carats
per cubic yard in angular quartz gravel. The diamonds resembled those at Gandorhun, but the average
size appeared to be much lower. A search of tailings dumps produced one rounded pebble. Heavy
mineral concentrates were sparse but consisted mainly of water worn corundum.

The significance of the scattered swamp deposits described above is not year clear. Stracke
(1963) suggests that the Dambu deposit must have originated by the reconcentration of Moa high
terrace material. Such high terraces have not been detected, but it is certainly ture that all the deposits
contain traces of alluvial material. However, it is most difficult to believe that the Moa river could
have supplied the diamonds in any of these deposits, in view of the extreme scarcity of diamonds in the
present Moa channel between the Guinea border and the Male confluence. Furthermore, the Teyama
deposit is a very long way from the Moa. It is more probable that all these swamp deposits are derived
frm scattered minor kimberlite outcrops.

Block 28 – MIDDLE MOA

This Block has been defined to include the channel, terraces and tributary swamps of the River
Moa for a distance of 20 miles downstream from the Male confluence. The Kenja stream (block 23)
discharges into this section of the Moa, which would thus appear to be most favourably situated from
the pont of view of alluvial diamond occurrence, since both the Kenja and the Male carry diamonds.
However, although there are diamonds in all the Middle Moa deposits, values are in fact extremely
disappointing. This is probably the result of dilution of the incoming diamondiferous gravels by the
overwhelming builk of the near-barren Moa gravels moving down from Block 27.
The Middle Moa flows across the Coastal Plain Surface with a mean profile gradient of 4 feet
per mile. The channel is in a stable semi-graded condition and has assumed a braided form, cotaining
numberous islands and islets but few rock-bars or rapids. The actual discharge is comparable to that of
the lower Sewa, but the channel is much broaded and shallower. All the Moa sediments contain far
more sand than the corresponding Sewa deposits, and the gravels tend to be finer and to contain a high
proportion of well-rounded quartz pebbles. These features of the gravel may possibly be the result of a
greater profile length, but in the absence of suitable maps of the Moa headwaters it is not possible to
decided whether or not the Moa is of a higher stream order than the Sewa. A contributory cause of the
gravel characteristics could be a deficiency of quartz veins in the local granitic basement; such a
deficiency has been noted in most exposures examined, but this may be fortuitous.

The only mining in the Moa channel has been in the form of sporadic diving, coupled with some
illicity pitting of isaldns and the adjacent shallow rocky areas. Recoveries have generally been rather
poor, the maximum obtained by the divers being 0.5 carats per cubic yard. The average width of the
channel is 1500 feet and there are few scoured sections; gravel yardages are thus very great but values
can only be guessed and are almost certainly at levels which will yield a small profit only to skilful
selective mining.

The Moa low terraces appear to be quite extensive, but some of the ground which has been
mapped as low terrace may in fact be flood-plain. They contain fine well-sorted gravels of bleached
rounded quartz, overlain by thick bleached sandy overburden with thin gravel lenses. Only a little
mining has been attempted by the diggers, and recoveries have never exceeded 0.2 carats per cubic
yard. Exploration of the low terrace deposits of the Maka area (418) was carried out by Minerals
Research Synidcate in 1962, but average recoveries were only about 0.1 carat per cubic yard. Very
substantial yardages of gravel certainly exist, but it is questionable whether any can be extracted at a

Only a few strips of high terrace gravel are known along the Moa, and values in them are very
low indeed’ no mining has been attempted. With one exception, the swamp deposits have all been
formed by the reconcentration of Moa low terrace gravels, and moderate recoveries have been made
from small selected patches of ground. The swamp gravels consist of bleached rounded quartz and
contain much water worn corundum. The exception is the swamp at Dia (417), where values of 0.25
carats per cubic yard occurred in angular quartz gravel below 1500 square yards of swamp. Rounded
pebbles were absent, but the corundum concentrates included a little water worn material. Pitting by
Minerals Research Synidcate established that diamonds were completely absent from other parts of the
same swamp.

The diamonds of Block 28 are clear octahedral of good quality and a median size of about 0.6
carat; coated stones are rarely seen. Heavy mineral concentrates in most deposits consist of water worn
corundum, chromate and zircon. Kimerlitic minerals were absent from all the Geological survey
samples, but the diamond Exploration Company reports the occurrence of grains of pyrope and
magnesian ilmenite in several samples from the Moa.

It has always been assumed that the diamonds of the Middle Moa were supplied by the Male and
proto-Kenja drainages, and the distribution and low average size of the diamonds lend support to this
assumption. There is nothing to suggest the presence of local sources, and although traces of kimerlitic
indicator minerals have been reported, these seem to be a universal feature of all the Moa gravels,
bearing no relationship to the presence or absence of diamonds and therefore of doubtful significance.

Block 29 – ZIMMI

The Zimmi Block includes a section of the Lower Moa and a group of scattered swamp deposits.
It is not covered by the available 1:50,000 maps and it is therefore necessary to refer to Map IV

The Moa here is 2000 feet wide; gravel reserves in the channel are consequently very large, but
average values are very low indeed. A limited amount o fmining is carried on, but it is mostly illicit
(recoveries being insufficient to cover licence fees) and take the forms of island and rock-bar pitting.
The most important swamps of the area are those at Kisiwuru and Bopo (425) which were first
discovered in 1963. there are two swamps, the Voblai and the Bunjiye, in both of which, values
between 0.3 and 0.7 carats per cubic yard occurred in thin sandy angular gravel. The Voblai
concentrates contained only angular material but the Bunife concentrates contained much water worn
corundum. The payable sections of these swamps are now mined out.

Between Zimmi and the Moa are several small low-grade swamp deposits in which intermittent
mining has taken place (426). Values in the angular gravels rarely exceed 0.2 carats per cubic yard,
and the average grade of the mining areas is about 0.1. To the west of the Moa are two more small
mined swamps near Gorahun which have not been examined but whose location has been noted from
the Diamond Exploration Company’s report on this area.

Very few diamonds from this Block have been examined, and there is therefore little that can be
said about them, except that only about 5% of the stones are coated. All the heavy mineral
concentrates contain corundum, rutile and zircon, but the most interesting feature of the Zimmi area is
the widespread occurrence of magnesian ilmenite. Concentrates from all the deposits contain a few
grains, but in the Bopo-Kisiwura deposits (425) this mineral is abundant in the 1mm concentrates. In
this region there can be little doubt that the swamp deposits, and probably also the Moa deposits, are
derived from local sources. Stracke (1963) infers the presence of a kimberlite zone west of the river,
and there must be another to the south-west of Zimmi and one near Kisiwura.

Block 30 – MORO

The east bank of the Moro river constritutes the international boundary with Liberia, and Block

30 covers those deposits which lie just of the Sierra Leone side of the boundary i.e. the Moro channel
deposits and the est bank terrace and swamp deposits. Most of the region, which is covered by high
rain forest, is sparsely inhabited and not very accessible; undetected illicit mining had been going on
for some time before attention was drawn to it in 1961 by N.W. Wilson, who was mapping the geology
of the area.

The Moro river is an 8th or 9th order stream which flows, with a gradient of 7.5 feet per mile,
along a valley floor which is an outlying tongue of the Coastal Plain Surface. The river is actively
incising this valley floor, so there are numerous rock-bars and rapids in the channel and frequent
narrow strips of terrace at various elevations. As far as is known at present, diamonds occur only
within the limits of the valley floor.

The Moro channel appears to contain diamonds throughout its entire length, but there is very
little information on which to base an estimate of values. Evidence of illicity mining can be found on
every rock-bar, but no diving or coffer-damming has yet been attempted. One diamond was found by
the Geological Survey on a rock-bar near Dambara (428). Recent flood=-plain rarely occurs, but there
is much low terrace which nobody has yet attempted to mine. High terrace deposits are extensive, but
again, no attempt has been made to mine them. Bulk sampling of the high terrace gravels in the Waima
area (427), by the Geological survey, produced no diamonds.

Swamp mining activity has been discovered at five locailities, and at all of them the swamps are
intersecting high terrace ground and the gravels contain reconcentrated terrace material. At Waima
(427), mining has taken place in three swamps. Sampling of virgin ground by the Geological Survey
produced average recoveries of 0.013 carats per cubic yard, but an examination of tailings indicated
that recoveries from the mined ground had been about ten times as great.

It has only been possible to examine a few Moro Valley diamonds, and these have all been clear
stones, many with inclusions and some with a brownish colour. Good actahedra appear to be
uncommon: most of the diamonds seen were flattened octahedral, dodecahedra or irregular pieces.
Heavy mineral concentrates from almost all deposits consist of corundum, ilmenite, rutile, zircon,
chrysoberyl and magnesian ilmenite. The magnesian ilmenite forms the bulk of the -2mm fraction in
some samples. Purple-pink garnet is present in all samples from the Moro channel but absent from all
terrace and swamp deposits.
Kimberlite exploration was carried out in the Waima area by the Geological Survey in 1963, and
this resulted in the discovery of four small dykes to the south-east of the village. Many more dykes
must exist in the vicinity. The dyke material, a mottled dark-brown clay, yielded magnesian ilmenite
and one diamond of 0.2 carat from 8 cubic yards. On the basis of these results, the dyke material was
originally reported as weathered kimberlite, but this identification is treated with some caution by
Wilson (1965). On further consideration of the texture of the clay, which contains no recognisable
relict kimberlitic textures, its identification as kimberlite is withdrawn and it will in future be referred
to nly as diamond source rock. The grade of the sample taken was 0.025 carats per cubic yard but this
is almost certainly not typical, far too high in fact, since the very low-grade swamp gravels in the
vicinity contain forty times as much magnesian ilmenite as the cource rock itself.

The widespread distribution of maguesian ilmenite shows that the Moro diamonds are related to
a major zone of well-exposed source rock, much of which must lie within Liberia. The work of the
Geological Survey at Waima, together with the universally low alluvial values, make it appear most
probable that the source bodies contain such low values that they are of no economic interest.

Exploration of the river channel deposits may have some chance of finding payable alluvial
ground, but in view of the inaccessibility of the region and the possibility of boundary disputes, it
cannot be said to be a very attractice proposition.

Block 31 – MALEMA

Block 31 covers the Malema field, which was discovered by the diggers in early 1964 and is
named from the Chiefdom in which it is situated. This is the only completely new field discovered
since the inception of the alluvial diamond Mining Scheme, and it is the only field which still has
potential for expansion, as many of the adjacent areas have not yet been thoroughly explored. The
deposits here are of exceptional interest in connection with the cource-rock controversy, because,
although kimberlitici minerals are completely absent from the concentrates, yet the isolation of this
field from all others makes it impossible seriously to suggest a relationship with any distant source.

Most of the Malema deposits lie on a part of the 750 foot Surfacwe which is partly surrounded
by remnants of the 100 foot Surface. The field is bisected by the Longula river, which is a 5th order
stream and has a gentle profile gradient as far as Dambara and for 2 miles downstream. After this, a
series of rapids and falls begins, and the river drops 100 feet in 2½ miles, down to the Coastal Plain
Surface. All the deposits are in forested country with few villages and, prior to the diamond rush, the
effective road head was at Jojoima, some 10 miles from the deposits. The first discovery was made by
a group of itinerant miners after some months of fruitless prospecting in this region, and it is worth
noting that their persistence was a direct result of the high incidence of corundum throughout the area;
without this corundum, the deposits would have remained undiscovered for many more years.

The largest deposits are those of the Leuya stream (431 to 434), which is mined almost from its
source to within ½ mile of its confluence wi the Longula, a total distance of 12,000 feet. The upper
Leuya has a steep gradient, and short rocky sections virtually devoid of gravel alternate with wide
swampy basins where shallow overburden covers a clay-rich angular gravel. Good values have been
found in all these basins and they are still being intensively mined. The uppermost basin (431) carried
values of 0.5 carats per cubic yard and was the first to be mined. Above it, the stream flows in a very
narrow valley and although many licences have bene issued up here, few diamonds have been found.
Downstream, recoveries varying from 0.3 to 1.2 carats per cubic yard have been made from the series
of basin deposits. The heavy mineral concentrates contain abundant corundum, both angular and
waterworn, and the distribution of the waterworn component corresponds closely with that of the
diamonds, indicating that the present diamond distribution has been determinted wholly or partly by an
ancient drainage pattern.

The lower Leuya (433, 434), which is a 3rd order stream, has a shallow gradient and flows
through continuous wide flats. On examination, most of the flats are found to be the deposits of an
earlier, larger stream.
Over burden is 8 feet in thickness and the gravels contain occasional rounded pebbles. Values in
the lower Leuya vary from 0.1 to 0.7 carats per cubic yard, and are accompanied by corresponding
concentrations of water worn corundum. At the time of the examination (April 1965) there was no
mining within a half-mile of the Longula confluence. Later in 1965, mining began in the upper Pume
stream, to the west of the Leuya. There are as yet no data available on this deposit, except that 31
licences were issued in 1965.

Another important deposit, now almost worked out, is the upper Kumiyama swamp (438, 439)
which is 8000 feet to the south-east of Dambara. Here, values between 0.4 and 1.0 carats per cubic
yard were usual, but some patches of ground in the swamp centre yielded 2.0 to 3.0 carats per cubic
yard. The gravel contained only angular quartz, but there was a little waterworn corundum in the
heavy mineral concentrates. This section of the swamp is upstream from the point where the main
Lasaru-Bandajuma path crosses. Downstream from this crossing is the lower Kumiyama (436) which
has nearly all been demarcated for mining licences but where there has been only intermittent actual
mining. The gravel here underlies a stream flat and contains much water worn corundum, but values
are low and much of it is unpayable for licensed miners.

7000 feet to the south-west of Dambara is the Keiya swamp (440), which also contained good
values and is now almost mined out. Mining has extended into a strip of high terrace along the north
side of the swamp and some of this remains to be mined. Both swamp and terrace contain only angular
quartz gravel, but waterworn corundum is abundant. Recoveries were highest along a zone following
the centre of the swamp. The presence of high terrace indicates that the swamp follows the course of a
former stream which presumably introduced the worn corundum and possibly also the diamonds. If the
stream came from the east, then the Keiya deposits may be related to those of the upper Kumiyama.

Finally, the Longula river itself. Since it bisects the field and all the mining areas are in its
tributaries, one might expect that the river deposits would themselves contain good vaues. So far,
however, this has not proved to be the case. A little mining has been attempted in the deep flats below
Dambara (435) but where gravel has been reached, only low recoveries have been made. Further
downstream, the Geological Survey carried out bulk sampling of gravel on a rock-bar (437), but no
diamonds were found.

Heavy mineral concentrates from the different parts of Malema field vary somewhat in quantity
but little in content. All contain corundum as a principal constitute exhibiting varying degrees of wear.
Minor constituents are ilmenite, zircon and rutile, with occasional abraded grains of red-brown garnet.
Small fragments of pink garnet occur in upper Kumiyama (438), apparently derived from nearby
granulites. In the Keiya (440), chromite is common and a little columbiteoccurs. No kimberlitic
indicator minerals were detected.

The Malema diamonds have a number of characteristic features. The majority are clear white
stones of very good colour and few inclusions. Pale greenish stones are found occasionally in all
deposits, but coated stones are rare to the north of the Longula and absent to the south. Good
octahedral are not common, butr a variety of distorted, crinked and flattened shapes may be seen. The
median size is about 1.0 carat and the size range seems to be unusually limited; the smallest stone seen
was 0.25 carat and the largest reported was one of 6 carats.

The iloation of Malema field from all other diamond fields can surely have no other explanation
than the presence of source bodies in the vicinity. If this is accepted, then the absence of the usual
kimberlitic indicator minerals from this field must be regarded s proof of the existence of a type of
kimberlite, or of some other source rock, which is deficient in indicator minerals. Such a rock is
thought to occur in many other districts where high diamond concentrations exist without kimberlitic
minerals, but only in Malema is no other explanation possible. There is much evidence in the Malema
deposits of the redistribution of diamonds within the field along ancient minor drainage routes, but the
grouping of the deposits shows that this process can only have operated over short distances. It seems
likely, for example, that the source of the Leuya and Pume deposits will be found a short distance to the
north, possibly on the broad forested wind-gap at the head of the Pume.

It is possible that future prospecting may extend the limits of the field, particularly in the
forested areas to the south, south=-east and east, where few people normally go.


The town of Mongeri is situated on the Teye River, 27 miles west of the Upper Sewa, and the
people of the town claim to have found one or two diamonds in the Teye channel gravels beside the
town. A Geological Survey prospecting party were unable to find diamonds in the rock-bar gravels or
in the terrace gravels, but did recover a 1.4 carat clear diamond from 11 cubic yards of gravel in a
swamp intersecting the second high terrace about 1½ miles to the south of Mongeri. After this
discovery, licensed mining began in the area, but it ceased after a few months, probably because of
poor recoveries. The Geological Survey continued prospecting upstream as far as Gbewama, but no
more diamonds were found. However, this work was overtaken by an early rainy season, consequently
the channel gravels could not be tested and sampling was limited to the terraces and tributaries. The
work could usefully be repeated, washing bulk samples from rock bars.

Selu is 10 miles to the east-south-east of Mongeri and is situated near the Koye River, a 7th order
tributary of the Teye. Illicit mining has occurred beside the Koye in a tiny swamp intersecting the high
terrace, but this swamp was cleaned out some years ago, and the Geological Survey party were unable
to find sufficient virgin gravel for a sample. High water levels prevented sampling of the Koye channel
gravels, but an abandoned channel of the river to the north of Selu was tested and one 0.5 carat clear
diamond was recovered from 5½ cubic yards of selected gravel. Subsequent prospecting of the entire
Koye drainage upstream failed to find any more diamonds.


Prospecting licences have several times been issued for the flats of the Kendi stream near Galu,
which is 16 miles to the north-east of Bo, and a few diamonds have been found. Attempts at mining,
however, have always been short-lived, because of very poor recoveries.


One diamond was found near the confluence of the Sengeye stream with the upper Tabe river,

13 miles north-west of Bo, in the course of Sierra Leone Selection Trust reconnaissance prospecting


Diamonds were found in the lower Waanje River near Baiama, and five miles to the east, in the
Makanu River, during S.L.S.T. reconnaissance prospecting work in the Pujehun area.


The Masau River runs parallel to, and a few miles to the east of the lower Moa. S.L.S.T.
prospectors found one diamond in the river near Foindu, and the Diamond Exploration Company found
one diamond in a pit near Bumpe in the course of their kimberlite prospecting.


In the extreme north of Sierra Leone, three diamonds were found by S.L.S.T. prospectors in the
Mongo River, above Musaia.

In the west of the country, a few miles to the north-east of Kambia, diamonds were found in the
Kankana river and in some of its tributary streams by S.L.S.T. prospecting teams. Only very low
recoveries were made, and the diamonds are said to have all been small in size and green in colour.

For many years, nothing was known of the source of the Sierra Leone alluvial diamonds.
Investigations began after Junner (1946) had postulated a principal source in the Woyie valley near
Koidu, because of the diamond and pyrope concentrations in that area. Early in 1948, the Sierra Leone
Selection Trust discovered kimberlite dykes in the Koidu area and in the Oyie valley, some 12 miles
distsant. Subsequent work resulted in the discovery of numerous other dykes, disposed in a series of
parallel zones.

At present, two widely separated groups of outcrops are known, the Koidu Dyke Zones and the
Tongo Dyke Zones, and both groups are described below. The Panguma dykes are part of the Tongo
group, but are the only ones outside the control of Sierra Leone Selection Trust; they have been the
subject of an investigation by the Geological Survey and will therefore be separetly discussed.

Although the diamond fields are very extensive, searches have failed to discover any other true
kimbelrites outside the two small groups of Koidu and tongo. Many people therefore believe that all
the alluvial diamonds throughout the field must have originated from one or other of these known
source areas. It is contended here, however, that numberous local sources exist which must differ from
normal kimberlite in appearance and heavy mineralogy. A full discussion of the source problem
follows the description of the kimbelrites.


(i) THE DYKES (Map Sheets 58 and 59)

A detailed description of the Koidu kimberlites has been given by Grantham and Allen (1960)
who showed that these rocks were correctly named, and some of the features described here have
already been noted in their paper.

The kimberlite occurs principally in the form of narrow dykes which intrude the granite; all have
a near-vertical dip, and their strikes all lie between 60° and 70° true. Individual dykes are short, in
most cases less than 1000 feet in length, but they are numerous and are disposed in a series of parallel
zones within an area 14 miles by 6 miles. All the known outcrops in this region are within the
Yengema field Lease of Sierra Leone Selection Trust. Each dyke is a composite body consisting of a
series of anastomosing thin dykes and stringers, the aggregate widths of which vary along the strike,
from a few inches to several feet. The structure of the kimberlite is in fact closer to that of a vein,
rather than to what is normally understood by the term “dyke”.

Immediatley to the south of Koidu Town an exceptionally dense swarm of dykes crops out, and
it is here that values are highest and dyke widths greatest, the average aggregate width being about 2
feet. A detailed sampling programme carried out here by the company has shown that the average
gradeof the weathered dyke kimberlite is 0.87 carats per cubic yard. This figure, however, is by itself
of little significance since each dyke in fact consists of several sub-parallel flat bodies interleaved with
barren granite, and all must perforce be mined and treated together. Calculations have therefore to be
made on the basis of values over the mining width. This is the minimum stoping width necessary at
any point to include all branches of the dyke, with an overriding minimum value of 3 feet. The
averagemining width of the Koidu dykes is about 4 feet, with an average mining grade of 0.46 carats
per cubic yard.

All dykes contain porphyritic kimberlite, normally without inclusions although the occasional
dyke does include a scatter of small granite xenoliths. Near the surface, dyke kimberlite weathers to a
greenish-brown clay to depths of 30 feet or more, and is far more susceptible to weathering than the
enclosing granite. It has been possible to study the unweathered rock in a dyke which is exposed in
underground workings near Koidu, at a depth of 300 feet. The dyke was found to be composite and to
contain different types of kimberlite. The commonest rock type was a blue-green porphyritic
kimbelrite; less common were a dark greenish-black rock with abundant small green phenocrysts, and a
dark grey fine-grained rock somewhat resembling limestone.
The petrography of the different rock-types may be summarised as follows:-

Normal porphyritic kimberlite (Specimen K22)

Abundant closely-packed oval olivine phenocrysts, partially serpentinised, in a fine-grained
matrix of phlogopite, calcite, serpentine and magnetite. Scattered books of coarse phlogopite.
Rare perovskite. One large rounded mass of ilmenite, with corroded margins. One irregular
garnet grain, enclosed by phlogopite, both enclosed by olivine.

Black olivine-rich kimberlite (Specimen K25)

Abundant small rounded olivine phenocrysts, mostly aligned parallel to the dyke walls. Very
slight serpentinizatrion only, along cracks. One broken olivine crystal, with separated
fragments. Groundmass of magnetite, calcite and a trace of serpentine. Very rare phlogopite.

Grey carbonate-rich “kimberlite”. (Specimen K26)

An aggregate of dine-grained carbonate, phlogopite and magnetite. Small clumps of phlogopite,
particlaly chloritized. Small oval granular carbonate pseudomiorphs olivine. This is not, strictly
speaking, a kimberlite.

Many intriguing relationships are disclosed by thin-sections of the kimberlites and of the
contacts with the wallrock. Among them are:-

1) The margins of many coarse ilmenite grains were undergoing corrosion by the matrix
immediately prior to final consolidation. Under high magnification, the edges are seen to be ragged
with an aura of detached gragments dispersing outwards.

2) The grey carbonate rock was the last to be emplaced, as it cuts the dykes of the other two rock-
types. In some cases however, the carbonate appears to have replaced the kimberlites rather than
intruded them; the progressive replacement of the olivines can be seen in many sections.

3) The contacts of the kimberlite with the hornblende-granodiorite wallrock are typically sharp and
irregular. There is often no evidence of movement, and only a little of alteration and replacement. The
actual contact is generally marked by a film of carbonate, and carbonate stringers occasionally
penetrate the granite. These stringers often cut through biotite or feldspar crystals, and in many cases
the veinlet/host relationship shows clearly that although the carbonate was originally injected along a
crack, subsequent thickening of the veinlet has been by replacement of the walls rather than by dilation
of the crack.

It is difficult at present to visualise the conditions under which these kimberlite dykes were
emplaced. The absence of wallrock alternation shows that the temperature of emplacement was low;
this fact and the presence of abundant rounded phenocrysts require the kimberlite to have been intruded
as a much rather than as a magma. Ultra high pressure must obviously have been necessary to force
this viscous material through narrow irregular fractures for several thousands of feet.

The anastromosing dyke systems generally appear to have followed pre-existing parallel
bifurcating fractures of shear zones. These could not have been open prior to intrusion, otherwise
sheets and slabs of granite between the fractures would have collapsed and the dyke would be filled
with coarse breccia and disoriented granite slabs. This is not known to occur.

It therefore follows that the kimberlite must have dilated the fracture as it rose. However this
was done, it is clear that the carbonate component of the kimberlite has played an important part in its
emplacement, since it is apparently capable at high pressures of replacing silicates, and it usually lines
the contact with the country rock. Drilling at Panguma has shown that at their extremities fine
kimberlite stringers narrow to become calcite veinlets.
(ii) (a) NUMBER 1 PIPE

In the Koidu area there are in addition to the dykes, four small pipes which are related to the
dykes. Number 1 Pipe crops out 4000 feet south of Koidu and is a tgrue pipe, filled largely with
kimberlite breccia. Although it lies near a kimberlite dyke zone, there is no doubt that it is a separate
body, not a dyke enloargement. At the surface it is very roughly oval in shape, with a maximum
diameter of 230 feet and a surface area of about 28,000 square feet.

In order to begin the underground development of Pipes 1 and 2 and the adjacent dyke zones,
Sieera Leone Selection Trust have sunk a shaft to a depth of 300 feet; Number 1 Pipe has already been
intersected by a drive at this level and bulk-sampling has been carried out by means of a series of
crosscuts. The pipe appears to plunge at 78° to the south, and at the 300 feet level it still has a cross-
sectional area of about 28,000 feet, although the cross-sectional outline has become rather rectangular.

98 feet to the south-east of the pipe margins, the drive passes through an 8 foot thick dyke of kimerlite
breccia which appears to dip towards the pipe and is probably an offshoot from it. No trace of this
offshoot has been found at the surface. 2200 carats of diamond were recovered from the first 1800
cubic yards of pipe breccia to be milled, giving an average grade of 1.22 carats per cubic yard or about

0.07 carats per long ton. This is, rather unexpectedly, only alittle lower than the values that were found
during bulk sampling of the decomposed near-surface material between the 12 foot and the 24 foot

The polymict rock which fills the greater park of the pipe is by convention termed a kimbelrite
breccia but it is in fact misnamed and could be more accurately described as a volcanic conglomerate.
The xenoliths, which make up 60% to 70% by volume of this rock, vary considerably in size bu the
majority are between 1 inch and 3 inches across. The matrix consists of a grey-green kimberlite
adulterated by pink felspathic debris and containing abundant rounded phenocrysts; serpentine
psedomorphs after olivine, oval phlogopite books, magnesian ilmenite and pyrope. The fine
groundmass consists of carbonate, serpentine and magnetite, with a little phlogopite. Olivine has not
been seen in the pipe breccia.

Representatives of many rock-types are found maong the xenoliths, and some of them pose
awkward problems. The following have been noted so far:-

Granites and granodiorites

These are generally pink hornblende-rich varieties, but the xenoliths are often stained green and
desilicated at the margins. At the 300 foot level they constitute about half of the total xenoliths,
but near the surface they are far more abundant. Some resemble the granites that enclose the

Hornblende rocks

A very wide variety of hornblende-rich rocks have been noted, ranging from pure hornblendite
to amphibolite. Similar rocks are of common occurrence as lenses and partically granitised
inclusions within the granites of the Koidu area.


Basalt xenoliths, sometimes several inches across, are common at the 300 foot level but rare near
the surface. The basalt is a fine-grained grey-black rock consisting of an ophitic aggregate of
plagioclase and augite, with scattered pyrite cubes. Vesicles contain an unidentified dark green
mineral and aggregates of very small crystals that may be either leucite or analcime. This would
appear to be a rock which originally consolidated from a surface flow; its appearance in the
breccia can only be explained by assuming that the pipe reached the surface at the time of
intrusion and that on this surface was a basalt flow of pre-kimberlite age. Disrupted basalt
became incorporated in the pipe and in some way managed to descent several hundred feet.
Basalts do not occur on the present surface in the Koidu area, but there are numerous dolerite
dykes which could well have been the feeders of basalt flows on older planation surfaces.
The suggestion will no doubt be made that the rocks which have been identified as basalts could
in fact be the chilled margins of dolerite intrusions cut by lower levels of the pipe. This appears
at first glance to be a credible explanation, but its acceptance is prevented by the complete
absence, among the xenoliths, of normal dolerite, which one would expect to be present in
excess of the chilled material.


Numerous small xenoliths of a brownish-black micaceous variety of kimberlite have been noted
in the breccia. These must represent earlier kimberlites which were intersected and disrupted by
the pipe.


Kikmberlitic Tuff

Fragments of a fine-grained purple-grey or purple-brown rock of shaly appearance are sparingly
distributed throughout the breccia. Some of them have a layered or bedded structure, and they
were originally thought to have been derived from formations of the Rokell River Series*, which
they resemble. Thin-section examination shows, however, that the rock consists of small grains
of carbonate, feldspar and serpentine in a very fine-grained brown matrix. A little phlogopite
can be detected in the matrix, but most of it is unidentifiable. This rock is now considered to be
a tuff, formed by explosive comminution of the pipe filling, and deposited in the original crater.

Carbonate grit

Rare and fascinating inclusions are those which resemble pieces of sandstone. One specimen
has been obtained and has been found to consist principally of rounded, sub-angular and angular
grains of carbonate. Some are fragments of larger crystals, others are microcrystalline. The
granular material encloses parallel flat lenses of dine-grained grey-green material which appears
to consist of comminuted carbonate and serpentine, and similar material is present in the
interstices between the granules. Flakes of phlogopite are present in the lenses, and are oriented
parallel to them. This would appear to be a water-laid deposit of kimberlitic material and is
therefore also classified as a tuff.


Eclogite inclusions are rare. The single specimen examined by the author consisted of pink
garnet and green much-altered clinopyroxene, and was cut by veinlets of carbonate. Grantham
and allen describe a similar example, and suggest that the garnet is a pyrope-almandine and that
the pyroxene is a dioside-jadeite. Eclogites are not known to occur elsewhere in Sierra Leone
(except in other kimberlite pipes).

* F.H. Hubbard has reported the occurrent of unmetamorphosed sedimentary rocks of volcanic origina
as xenoliths in the pipe. (Nature, 1967 Vo. 214 No. 5092, pp 1004-1005).

Garnetiferous serpentine

This rock has not been found during the present investigation but is recorded by Grantham and
allen, who list its constituents as pyrope, serpentine, streaks of iron ore, and veinlets of
carbonate. It does not occur elsewhere in Sierra Leone.

Along certain sections of the pipe walls, but still within the pipe are large irregular masses and
pockets of porphyritic kimberlite without xenoliths. This is greenish-black, micaceous, riddled with
carbonate veinlets and exceptionally susceptible to weathering; it would appear to represent the final
phase of intrusive activity. At the pipe margins, the pipe filling is often separated from the wallrock by
a lining of dull green or red-brown asbestiform material, about half-an-inch in thickness and consisting
of talk and calcite.
The country-rock surrounding Number 1 pipe is a coarse-grained hornblende granodiorite which
is sliced through by flat-lying water-bearing joint planes. In the section approaching the pipe, the main
drive on the 300 foot level follows a flat-lying body of basic rock within the granite for over 100 feet.
The rock is medium-grained, grey to black in colour and has sometimes a schistose and sometimes a
hornfelsic texture. It consists of green hornblende biotite, labradorite and a little sericite, and in many
specifmens there is parallel orientation of the grains and vestigial foliation. Light spots visible in the
hand specimen are found to consist of labradorite clusters with augite around the edges. The
occurrence of this rock as a persistent flat sheet precludes the possibility that it is an inclusion of a pre-
granite formation, although it is cut in places by veins of granitic composition. It is now a
metamorphic rock, an amphibolite or epidiorite, and probably originated as a basic sill which intruded
the granite. Vertical dykes of a similar nature, apparently meta-dolerites, have been noted elsewhere in
the diamond fields.

(ii) (b) NUMBER 2 PIPE

The existence of Number 2 Pipe, half-a-mile to the east of Number 1, has been known for some
time but the detailed investigation has only recently begun. Number 2 Pipe was originally thought to
have an outcrop area of about 30,000 square feet but it now appears that its actual extent may be two or
three times this figure.

Four distinct types of kimberlitic material occur in the pipe:

1) Porphyritic kimberlite, in the form of irregular masses, dykes, and possibly sills, all connected to
a network of dykes which cut the adjacent granite.

2) Polymict kimberlite breccia, similar to that which forms the bulk of Number 1 Pipe. Abundant
xenoliths of a wide variety of rock-types in a predominantly kimberlitic matrix.

3) Fine felspathic kimberlite breccia, which consists principally of fine granitic debris intimately
mixed with a little kimberlitic material. The term “tuffisite” appears to be applicable, and is certainly
less cumbersome, and will be used throughout the following discussion.

4) Boulder breccia, which consists of close-packed, well-rounded granite boulders up to 3 feet in
diameter in a tuffisite matrix. The boulders consist either of pink porphyroblastic granodiorite,
identical with the pipe country rock, or of amphibolite, which is common in the form of inclusions
within the country rock.

All these four components of the pipe contain diamonds, and values appear to be highester in the
porphyritic kimberlite.

The portion of the pipe now exposed originally formed the bedrock of a shallow stream-flat, and
the delineation of the pipe margins in the adjacent slopes is now proceeding. This is by no means
straightforward since both the tuffisite and the boulder breccia weather to mottled lateritic clays, with
ferruginous nodules, which are virtually indistinguishable from weathered granodiorite.

The pipe structure is complex, and the data available at present are too limited to allow the
relationship between the different components to be conclusively determined. Three masses of boulder
breccia lie on top of the other pipe components and are believed not to continue at depth. They would
appear to be basal remnants of a formerly continuous body, with a somewhat irregular base, which
formed a capping to the pipe. The close correspondence between the boulder rock and the country rock
shows that the boulders have not moved very far from their source level, and the lack of variety of
rock-types makes it difficult to believe that the boulder breccia represents erupted material which has
fallen back into the crater. It seems more likely that it represents the top level of a pipe which failed to
blow out at the surface.
The boulder breccia is usually, but not invariably, separated from the underlying polymict
kimberlite breccia by a layer of tuffisite which may be up to 15 feet in thickness. Where the kimberlite
breccia has been exposed by removal of the overlying formations it can be seen that much of it has
been displaced by porpyhritic kimberlite. The core log of a vertical diamond drill hole, put down some
years ago into the centre of the pipe, shows that the pipe at depth consists principally of polymict
kimberlite breccia which contains a few large granite blocks and is cut by several bodies of porphyritic

The granodiorite country rock adjacent to the pipe outcrop is intersected by numerous dykes and
stringers of porphyritic kimberlite, most of them with their strikes approximating to the usual
kimberlite strike, 65° true. It is clear that the emplacement of these dykes took place after
emplacement of the pipe breccias, since some of the dykes cut through the breccias and tuffisite and are
connected to the large irregular masses of porphyritic kimberlite which displace much of the breccia in
the present pipe exposures.

The injection of porphyritic kimberlite seems to have occurred as the final phase of an intrusive
cycle, and is possibly related to consolidation and subsidence within the quiescent degassed pipe. Ion
the case of this particular pipe, the field evidence indicates that the kimberlite dykes, injected into the
fractured granite during this final phase, penetrated to much higher levels than the pipe itself.

Two important practical issues are raised by the above interpretation of the Number 2 Pipe
exposures, and these are:

1) The kimberlite dykes in the vicinity of the pipe are clearly related to it. Their emplacement was
apparently guided by the pipe itself and by structures produced in the granites by pipe intrusion and
subsidence. Possibly many kimberlite dykes have a similar history; this would accord well with their
usual mode of occurrence as groups of short discontinuous dykes. The 65° fractures in which the
majority of dykes are emplaced are of local importance only, not deep regional faults; this is borne out
by the underground exposures which show that relative displacement of dyke walls is only one or two
feet and that cataclasites or fault beccias are absent. If the suggested pipe control of dyke emplacement
is a principal of general apoplication, then any group of kimberlite dykes indicates the presence of a
pipe which, if not detected at the surface, should be sought at depth.

2) The boulder breccia and the tuffisite, when weathered, form mottled lateritic clays not easily
distinguishable from those formed from normal granodiorite. If a pipe outcrop should consist only of
these formations, it will have been excusably overlooked in the past and will not be easy to detect in the
future. Such bodies may well account for some of the otherwise unexplained alluvial diamond
concentrations in Yengema Field.

(ii) (c) OTHER PIPES

Number 3 Pipe crops out one-and-a-half miles to the east of Koidu town and lies under a swamp.
It has not yet been investigated in detail but it appears at present to be smaller than Pipes 1 and 2 and to
consist of both kimberlite breccia and porphyritic kimberlite.

A fourth pipe occurs almost midway between Pipes 1 and 2. it is very small, having a surface
area of only 5,500 square feet, and because it is located on a kimberlite dyke it has always been
regarded as a dyke enlargement. Underground work has now shown that it is a true pipe of polymict
kimberlite breccia and that it cuts and displaces the kimberlite dyke.


The age of intrusion of the Koidu kimberlites is not yet known. J.F. Lovering and W. Compston
(in preparation) give the Rb/Sr. age of phlogopite, separated from the kimberlite of Number 1 Pipe, as
about 170 million years, and this of course refers to the time elapsed since the strontium in the
phlogopite was isolated from the strontium in the kimberlite matrix.
Most of the phlogopite in the pipe kimberlite occurs in the form of small oval books which must be
regarded as phenocrysts formed before intrusion. The date obtained, therefore, is a maximum age for
the kimberlite; the actual date of emplacement and final consolidation could be much later. All lthat
can be said at the moment regarding the age of the kimberlite intrusions is that they are almost certainly


The kimberlites of the Tongo group crop out in the Tongo field Mining Lease of Sierra Leone
Selection Trust (Map Sheet 81), 30 miles due south of the Koidu group. The tongo kimberlite dykes
compare with the Koidu dykes in many ways; they have the same steep dip, 60° to 70° strike and
anastomosing structure, but they are fewer in number and generally much narrower. Values in the
kimberlite, however, tend to be higher than at Koidu. Most of the Tongo kimberlites contain abundant
magnesian ilmenite and chromite, but only a little pyrope garnet. One very small pipe has been found.

Heavy mineral sampling grids have outlined four dyke zones, each containing several short
dykes. The most important is the Lando Zone, which is almost 5 miles in length and skims along the
south side of the Lando stream. Trenching of the central part of this zone has shown that the average
dyke width is 13 inches over a strike length of 8700 feet, and these small dimensions are thought to be
sub-economic for an underground mine. Nothing is yet known of the petrology of the dykes, as only
weathered material is available for examinsation, but thepresence of oval spots resembling the usual
olivine pseudomorphs, and of magnesian ilmenite and diamond, leave little doubt that they are in fact

Some work has also been done on the Peyima Zone, which is about 3 miles in length and lies to
the south-west of Laoma. Within this zone, the average dyke width is about 9 inches and values are
comparatively low. The main feature of interest here is the small beccia pipe which measures 20 feet
by 50 feet at the outcop and cuts the principal dyke beside the Peyima stream. This body is composed
of a coarse kimberlite breccia with an average diamond content, in the weathered zone, of 0.83 carats
per cubic yard. The beccia matrix has not been seen in an unweathered condition and cannot therefore
be accurately described, but in one corner of the sampling trench some semi-weathered material was
visible and in this the matrix appeared to be a bluish porphyritic kimberlite. It bore more resemblance
to the rather micaceous dyke kimberlite of Panguma than to the grey-green feldspar-adultrated matrix
of the Koidu breccias. Xenolithic material in the sample trench was in various stages of
decomposition, but identifiable examples of the following were obtained.

Coarse biotite-schist
Garnetiferous basic granulite

The first five rock-types are common in the Tongo region, but the last two are not.

The garnetiferous basic granulite is a medium-grained olive-brown rock with a compact granular
appearance. In this-section, it is seen to consist of equigranular plagioclase with a mosaic texture, and
scattered grains of euhedral pink garnet. Traces of zircon, apatite, ilmenite, pyroxenes and secondary
amphiboles are occasionally present. Although rocks of the granulite facies are common as strips
within the granite of south-eastern Sieera Leone (Wilson, 1965, and Andrews-Jones, 1966), this
particular rock-type has not been recorded before. The eclogite inclusions are rare. They consist of
closely-packed pale pink euhedral garnets, with the interstices filled with greenish much-
alteredclinopyroxene. The only other known occurrence of eclogite in Sierra Leone is in the kimberlite
breccias of the Koidu pipes.

The kimberlite dykes of the Upper Talama area (348), near Panguma, occur within a zone which
is 1800 feet wide and approximately 6000 feet long. This zone is apparently unconnected to any other
zone outcrop, but it is part of the Tongo kimberlite group and could be related to the Lando zone at
depth. The Panguma zone is the most westerly member of the Tongo group and it is terminated in the
west by the Maboa fault, although only one dyke outcrop actually meets the fault.

The dykes were discovered in 1960 by the Geological Survey. A detailed examination was
carried out in 1961 and 1962, and this established the surface characteristics of the zone. In 1964 and
1965, a programme of diamond drilling was undertaken in order to determine the nature and
dimensions of the kimberlite bodies at depth.

At the surface, 7 separate dykes were found, with lengths varying from 200 feet to 1000 feet.
The strike of the zone is 65° true but the strike of the individual dykes may vary by 15° on either side
of this figure. Dips are always between 80° and 85° to the north. Each dyke actually consists of an
anastomosing network of minor dykes and stringers, and the average aggregate thickness of these
components is 18 inches. In addition to the dykes, there are a few outlying stringers, one inch wide or
less, some of which were noted in trench exposures and some of which are inferred from heavy mineral

Sampling of the weathered kimberlite gave an average recovery of 0.77 carats per cubic yard,
but this figure must not be regarded as final. Sampling was handicapped by the activities of illicit
miners, consequently only 11½ cubic yards were treated and this is not really an adequate sample.
Decomposition of the kimberlite by weathering extends to a depth of 20 to 30 feet and all the dyke
outcrops have now been illicitly mined almost to the weathering limit. The narrow stope-like trenches
made by the miners have all subsequently collapsed, and it is therefore no longer possible to inspect
kimberlite exposures unless a major excavation is made; there is in any case very little material left
which is sufficiently decomposed to be suitable for sampling and treatment in field equipment.

In the course of the sub-surface exploration, 8 inclined holes were drilled, their lengths varying
from 350 feet to 1000 feet. 8 dyke intersections ere made. The best was a 27 inch single dyke, thicker
than the corresponding outcrop; each of the others was a collection of very thin dykes and stringers
with aggregate widths either equivalent to or less than those of the corresponding outcrops. A few thin
stringers were met which had not been detected at the surface.

One hole was drilled through the line of intersection of a kimberlite dyke with the Maboa fault
zone, after attempts to expose the outcrop had failed because of the great over-lying thickness of
waterlogged tailings. This hole passed through a 400 foot zone of fractured sheared and occasionally
brecciated granite, but there was no sign of kimberlite. It seems improbable that the rising kimberlite
could have met this zone of shattered rocks without invading it, and it would therefore appear that the
principal movements along this fault took place after the intrusion of the kimberlite.

The dyke rock is a greenish-black, porphyritic, visibly micaceous kimberlite. It contains large
oval olivine phenocrysts, partly serpentinised, in a matrix of phlogopite and carbonate, with a little
magnetite. In some of the thinner dykes the olivine is completely serpentinised, and in a few cases the
serpentine pseudomorphs are undergoing marginal replacement by carbonate. Occasional rounded
grains of magnesian ilmenite are visible in the hand specimen, but pyrope has not been seen.

In most dykes, the contact with the country-rock is marked by dark-green fibrous carbonate, and
very often the inner surface of the carbonate film is slickensided. The granite country-rock close to the
dykes is frequently reddened and chloritized, and may be cut by fine stringers of green carbonate and
phlogopite. These effects are most pronounced around and among networks of thin dykes but are very
slight in the vicinityh of the thicksingle dyke.


(Map Sheet 81)
Isolated stringers with a thickness of 2 inches or less show a pronounced decrease in the size and
quantity of olivine phenocrysts and an increase in the carbonate constituent. Although the layout of the
drillholes did not permit any individual stringer to be followed, all gradations have been noted in the
cores, from ¼ inch stringers of carbonate-rich kimberlite to ⅛ inch veinlets of green carbonate with a
trace of phlogopite and 1/16th inch veinlets of pure carbonate. Within a few feet of each dyke there are
generally numerous tight fracture planes, parallel to the dyke walls and lined either with carbonate or
chlorite. It is not clear whether the fracturing preceded or accompanied the kimberlite emplacement.

In all the Panguma drill-cores the country-rock, whether adjacent to a dyke or not, is cut every
twenty or thirty feet by a fracture lined with a carbonate, generally calcite but occasionally ankerite.
The country-rock is a microline granite, with biotite-rich zones and frequent amphibolite bands, and it
seems unlikely that the carbonate is a normal feature of these rocks, even though cores of
unmineralized granite are not available for comparison. It will probably be found that widespread
carbonate impregnation is a characteristic feature of zones of kimberlite emplacement.

The conclusions drawn from the surface and sub-surface work done at Panguma were as

(i) The dykes consist of porphyritic kimberlite. Kimberlite breccia does not occur and there are no
pipes or substantial dyke enlargements within 600 feet of the surface.

(ii) The total strike length of dyke exposed at the surface, outside the S.L.S.T. Mining Lease, is 4200
feet and the average dyke width is 18 inches. Values must provisionally be assumed to correspond
with the sample recovery i.e. 0.77 carats per cubic yard.

(iii) At the depths drilled (400 to 700 feet), there is no evidence of any general increase of dyke
thickness with depth.

(iv) In addition to the dykes there are several isolated kimberlite stringers within the zone, some of
which do not outcrop. These are all less than 3 inches in thickness, and whilst they may expand into
dykes at depth, this development is unlikely to take place at a depth of less than 1000 feet and they
must therefore be disregarded in initial mining calculations.

(v) In view of conclusions iii and iv, mining estimates must be made on the assumption that the only
bodies of interest are the outcropping dykes and that their dimensions and values continue unchanged
at depth.

(vi) The average minimum stopping width for the variable bifurcating dykes would be approximately
4 feet.

(vii) The following provisional estimates have been made from the available data:

Diamond resources in possibly workable bodies =180 carats per foot of depth.
= 180 carats from the surface to depth of 100 feet.

The estimated value of these resources is Le.3,600,000 (£1,800,000). Average mining width is 4
feet. Therefore in order to recover the stimated caratage it will be necessary to stop and treet
approximately 620,000 cubic yards of mixed kimberlite and granite from a number of separate flat
vertical bodies between the surface and 1000 feet. The average recovery from the mineral mined and
milled will thus be a 0.29 carats per cubic yard and the estimated revenue therefore Le.5.8 per cubic
yard or Le. 2.0 per ton.

Costs of mining and treatment are estimated at not less than Le.8.000 per ton.

(viii) The Panguma kimberlites are therefore not considered to be a viable mining proposition.


In the preceding sections, the diamond-bearing kimberlites of Koidu and Tongo have been
described. Both groups of kimberlites contain pyrope garnet and abundant magnesian ilmenite,
minerals which are relatively resistant to weathering processes and are readily detectable both in
kimberlitic eluvium and in alluvial concentrates within a mile or two of the parent kimberlite body.
Soon after the discovery of the kimberlites, it was found that the delimitation of kimberlite zones was
relatively straightforward when based on significant heavy mineral concentrations in soils and

Both groups of kimberlites are associated with rich alluvial diamond concentrations, but there
are also very extensive alluvial fields at some distance from the kimberlites. These fields were
originally outlined by the Sierra Leone Selection Trust reconnaissance prospecting work, in the course
of which two significant factgs emerged. The first was that kimberlitic minerals occurred in none of
the concentrates collected from fields which were distant from the known kimberlites. The second was
that almost all the alluvial diamond deposits in south-east Sierra Leone were connected to one of the
two known kimberlite areas, either by continuously diamonderriferous drainage (e.g. Sewa, Bafi,
Moinde, Meya) or by an ancient drainage route (e.g. Kenja, Beeya, Mamaye, Luya).

The inescapable conclusion was that all the alluvial diamonds in the fields had originated from
the Koidu and Tongo kimberlite groups and that they had been carried to their present situations by
ancient rivers. Some of these ancient rivers, such as the Bafi-Sewa system, appear to have persisted
with little modification into the present, but others have completely vanished.

This explanation of the diamond distribution became generally accepted, and in 1958 there was
no source problem since everyone concerned agreed that all alluvial fields could be satisfactorily
accounted for and that there was no prospect of finding anything other than trivial kimberlite
occurrences away from the Koidu and Tongo areas. In 1960, the validity of the alluvial theories was
questioned, first by Dr. b.B. Brock, and then by other geologists of the Diamond corporation who came
to investigate particular aspects of diamond mining on behalf of the Sierra Leone government.
However, after carrying out a search of the entire alluvial fields for kimberlite, these geologists also
concluded that there were no kimberlites of importance outside the Koidu and Tongo groups and they
accepted the hypothesis of alluvial dispersions.


At the beginning of the alluvial diamond survey by the Geological Survey, the trust of the
alluvial dispersion concept was acceptable as self-evident. However, as the survey proceeded, it
gradually became apparent that there are a number of serious objections to the proposition that all the
diamonds come from either Koidu or Tongo. Much attention has therefore been paid in this Bulletin to
the subject of diamond sources, as it is considered to be one of paramount economic importance.
Diamond sources are primary deposits which persist in depth and may be workable. If found and if
workable they provide reserves which can considerably extend the life of diamond mining. Acceptance
of a false hypothesis of diamond origins means that the source rocks will be sought in the wrong places
and may therefore not be found.

The general objections to the concept of alluvial dispersion are:

1) the kimberlite bodies which crop out in the known source areas are of insignificant
dimensions when compared with the total original allulvial diamond resources in south-
east Sierra Leone, which were about 64 million carats. This objection can of course be
countered by imagining the present kimberlites to be merely the roots of vast comfplexes,
now removed by erosion.
2) Whilst the transport of diamonds by streams with a suitably high gradient is not
disputed, the alluvial processes seen to be operating at present are manifestly inadequate
to have transferred a few million carats of diamonds from the Koidu to the Lower Sewa, a
distance of 120 miles, within any credible interval of geological time. This objection has
been countered by Haggard (1963 and 1965) who introduced the concept of “gravel
trains”. He suggests that at some time in the Tertiary, more arid conditions prevailed in
the diamond fields, resulting in the formation of builky rock gravels which formed
alluvial gravel trains capable of carrying large diamonds over ldong distances. He
maintains that rapid movement of the gravel trains towards the coast could have resulted
from Late-Tertiary tilting. There is, however, no evidence of tilting in the late-Tertiary.

3) In the course of the present investigation, many deposits have been found in which
evidence of alluvial transport is comfpletley lacking. Such deposits are particularly
common in parts of Blocks 13, 15, 17 and 18, in other words, those areas of the Coastal
Plain Surface which are often loosely but inaccurately referred to as the “Lower Sewa
Valley”. In these deposits there are no water-worn heavy minerals, no rounded quartz
pebbles and no topographical indications of any old valley.

4) Kimberlitic minerals are not, as was at first thought, completely absent from all
deposits at a distance from Koidu and tongo. Occasional small grains, principally of
magnesian ilmenite and all under 2mm, in size, have been recovered from deposits in all
parts of the diamond fields, both by the Geological Survey and by the diamond
Exploration Company. Stracke (1963 thinks that these grains were transported from
Koidue or tongo, with the diamonds. Apart from the basic improbability of the transport
and survival of ilmenite in detectable quantities over such distances in Sierra Leone
conditions, Stracke’s explanation cannot possibly apply in those deposits, fifty to a
hundred miles from the supposed sources, where the magnesian ilmenite is relatively
abundant. Examples are the Segyei flats (194), the Kokoye flats (215), and the Sewa low
terrace at Hima (310); all of these deposits are along the lower Sewa, about 100 miles
from the supposed sources, yet contain from 20 to 60 grains of -2mm magnesian ilmenite
per 100 c.c. of heavy mineral concentrate. This contrasts with the middle Sewa deposits,
much nearer to Koidu, where the mineral is virtually absent. These data indicate that
there must be rocks of kimberlite type in other parts of the fields beside Koidu and tongo.

5) There are numerous instances of deposits which, although part of an alluvial field
which is ostensibly related to an exotic source, nevertheless produce diamonds with
distinctive colour or shape characteristics. Examples can be found in all Blocks. Perhaps
the most striking are the Folu gravels (Block 26) and some of the Male gravels (block

25), in both of which between 10% and 15% of the diamonds are coated stones, yet they
are supposedly related to the tongo sources where coated stones are rare.

6) The situation of the Malema field (Block 31), discovered in 1964, is such that it can be
related to no known kimberlite zone. It is surrounded by barren country. Local source
rocks must be present, yet kimberlitic heavy minerals are completely absent. This
constitutes virtual proof of the hypothesis that kimberlites or other diamond source rocks
exist in which the normal indicator minerals are deficient or absent. If they exist here,
there is no reason why they should not exist in many other parts of the diamond fields
where there are neither kimberlitic nor convincing alluvial indications.


Source-rock investigations have been carried out by three orgnisations; Sierra Leone Selection
Trust, the Diamond Exploration Company and the Geological survey.
(a) S.L.S.T. kimberlite investigations of the normal kimberlites within their Lease areas by means of
soil sampling grids, pitting and tenching. The distribution of indicator minerals in soils has enabled the
Company to delineate a number of kimberlite zones in each Lease, but trenching within these zones has
disclosed only the narrow kimbelrite dykes and the four small pipes described in Section C. No major
kimberlite bodies have been found.

(b) The Diamond Exploration company kimberlite search

Between August 1961 and May 1963, geologists of the diamond Exploration Company, assisted
by twelve Sierra Leonean samplers, carried out a heavy mineral sampling programme over all parts of
south-easteren Sierra Leone not leased to S.L.S.T. Concentrates were collected from alluvial gravel at
intervals of about two miles throughout most of the drainage pattern and were sent to Johannesburg for
laboratory examination. Altogether, 4070 samples were collected. The sampling project and its results
are described in detail by Stracke (1963). The results may be summarised as follows:-

1) No indicator minerals whasoever were found in the Kenja/Beeya drainage (Blocks 22 and 23)
or the Matemu/Gbatiye drainage (Block 10), both of which contain diamond alluvials.

2) Small quantities of indicator minerals occur in occasional samples throughout the length of
the Sewa channel and in the terraces. Stracke states that they occur in diminishing quantities
proceeding downstream but this is not strictly true. Most samples in the Jaima, Jagbwema,
Konto and Barma areas (Blocks 4 to 6) contained some indicator minerals, and it is true to say
that there is a notable diminution proceeding downstream past Boajibu. However, further
downstream still, along the Lower Sewa, indicator minerals are found in an increasing
proportion of samples, not only from channel gravels but also from tributary swamps and

3) Indicator minterals were found in the Woa and Male gravels (Blocks 24 and 25).

4) Pyrope and magnesian ilmenite were found all along the Moa river, in the channel gravels.

5) Immediatley south of Yengema Lease, indicator minerals occur in most of the streams,

although diamonds are scarce.

6) Large concentrations of indicator minerals were found along the Moro river (Block 30).

The Diamond Exploration company undertook follow-up work at six localities. In the Luyei
stream and the Meni stream, both of them tributaries of the Waanje River with no alluvial mining
activity, anomalous concentrations of magnesian ilmenite were found. Large samples of gravel were
washed but diamonds were found to be absent. Detailed sampling of the Woama area, on the Woa
headwaters, showed that indicator minerals were abundant but diamonds were absent. On the Masaye
south of Njaiama (Block 3) indicator minerals were traced to their source, a thin discontinuous dyke
under the stream flats. This was thought to be kimberlite, but no diamonds were found. In a mined
swamp at Dambu, between Segbwema and the Moa (Block 27), a single grain of chrome diopsite was
found, but trenching showed that the bedrock consisted only of decomposed granitic rocks. A number
of anomalous samples between the Masau and the Lower Moa corresponded with some small illicit
diggings and were thought to indicate thepresence of small low-grade kimberlite bodies.

Stracke concludes that no kimberlites of economic importance exist in south-east Sierra Leone
except in the Koidu and Tongo source areas and that the great majority of Sierra Leone diamonds
originated from these areas. He suggests that recognisable indicator mineral grains, especially
magnesian ilmenite, can be carried in the drainage to distances of at least 150 miles from the source
and that the widespread occurrence of small quantities of indicator minerals along the Sewa, Male and
Moa rivers is a result of such alluvial dispersion from known source areas.

In June 1963 the diamond Exploration Company completed its kimberlite survey by taking large
(1½ lb) heavy mineral samples from major rivers throughout Sierra Leone at points where they are
intersected by roads. 26 samples were taken from a region of approximately 16,000 square miles.
None contained any kimberlitic minerals and it was concluded that there were therefore no kimberlite
bodies in the region. In view of the sample density, such a conclusion seems to be totally unwarranted.
(c) Geological Survey Investigations

Source-rock investigations have been carried out by the Geological Survey at a number of

(1) In 1960, 1961 and 1962, normal kimbelrite dykes were located at Panuma by using the
established technique of loaming for indicator minerals. These dykes were subsequently explored in
detail by diamond drilling and are fully described in the previous Section.

(2) In 1961, abundant magnesian ilmenite was found in heavy mineral concentrates taken from very
low-grade diamondiferouos alluvials along the Moro river, on the Liberian border. In 1963 the area
was explored for kimberlites, using the same loaming techniques as were employed at Panguma. Four
brown clay dykes were found which cut the decomposed granite under a swamp near Waima and which
yielded magnesian ilmenite. One diamond of 0.21 carats was recovered from this brown clay, and it
was therefore concluded that this was decomposed kimberlite, although it bore only partial resemblance
to the decomposed kimberlites seen elsewhere. It is now agreed that the identification is not justifieid
by the scanty data, and the dykes will henceforth be referred to as unidentified source rocks. Whatever
their true nature, however, it is almost certain that they are of no economic significance, since the
quantity of ilmenite in the alluvials shows that outcrops must be very numerous, yet alluvial diamond
values are extremely low.

(3) By the end of 1963, evidence had begun to accumulate which tended to show that source rocks
must occur throughout the diamond fields. Heavy mineral sampling both by the Geological Survey and
by the Diamond Exploration Company had already established the scarcity of indicator minerals in
most parts of the fields, and the normal methods of kimbelite search were therefore inapplicable.
During the 1964 and 1965, other techniques were tried by the Geological Survey.

After inspection of airphotos, a section of the mined-put flats of the Gbatiye stream near
Nongoba (Block 10) were selected for investigation. High values had originally been present in the
alluvial gravels. Banka drilling of the flats showed that the granite bedrock was cut by a number of
elongated bodies of dark greenish-brown mottled micaceous clay, the largest being 500 feet long and
150 feet wide. Deep trenches and cased pits were sunk to expose and sample these clay bodies, and
235 cubic yards of decomposed material were extracted and washed. The total yield was 13 diamonds
weighing 1.6 carats. This insignificant quantity must be attributed to penetration from the original
overlying alluvium, even though the top 18 inches of clay was discarded during sampling. The precise
nature of the bodies could not be established, but in some places the clay was rich in vermiculite and in
others there were rounded masses of talc. They would therefore appear to represent decomposed
ultrabasic rocks.

(4) After the failure of the Nongoba investigation, the Noniyei swamp near Kponima (Block 13)
was selectged for further work in 1965. Here in a swamp aligned at 68° ture, high alluvial values had
occurred and were associated with concentrates which contained rare grains of magnesian ilmenite and
some orange-red garnet. Early trenches showed nothing, except decomposed granite bedrock with
occasional micaceous quartzite inclusions. Banka Drill lines, with holes at 2½ foot intervals, were set
out across the swamp upstream from the trenches. A thin dolerite dyke lay beneath and along the axis
of the swamp, and the only item of interest disclosed by the drill cores was a 10 foot zone or dyke of
dark blue/green/balck plastic clay following the northern margin of the dolerite and in contact with it.
Cores of this clay were panned and yielded fine concentrates consisting principally of spheres of
siderite, withmuch magnetitie and diopsite, a little colourless zircon, and rare red garnet fragments.
Tenches were begun in order to sample this unusual clay but work was stopped by inundation of the
swamp during the rains and there was no further opportunity to continue it.

5) In the Sembehun 17/Nyandeyama area, also in Block 13, it had been noticed that the alluvial
diamond concentrations were all associated with concentrations of chrome spinel and actinolite. As a
similar relationship had been noticed in some other Blocks, it was decided to investigate it by means of
a loaming programme, which would not be handicapped by the rains.
In this area, which is part of the coastal Plain Surface, the diamondiferous swamps are separated by
broad flat watersheds on which thin patchy duricrust has formed. Examination of the concentrates
from the soil sampling grid enabled zones of high chrome spinal concentration to be delimited on the
watersheds. Several trenches were cut across these zones and 60 cubic yards of lateritic gravel and
decomposed bedrock were washed. One solitary diamond of 0.06 carat was found in lateritic gravel,
but additional bulk sampling in the vicinity failed to find more. The chomre spinel and actinolite were
traced to thick lenses of talk-tremolite schist which are enclosed by the granite, but the association of
these minerals with diamonds in certain alluvials appears to be fortuitous. The significance of the
single diamond is not known. It was several hundred feet from the nearest swamp deposit, and the
associated concentrates contained no trace of alluvial material.

The swamp and stream deposits of the Sembehun/Nyandeyama area have, in the past, often been
considered to mark the route of an ancient Sewa valley. If that were true, the present flat intervluves
would represent the former valley floor and should have some diamondiferous alluvial residues. In
view of our failure to find the diamond source rock, it was decided to check on the Sewa hypothesis
once again, in spite of the objections to it. A line of pits at 50 foot intervals was placed across one of
the flattest interfluves between two long swamps which formerly contained high values, and the
residual gravel from each pit was washed. There was not trace either of diamonds, or of rounded
pebbles, or of water worn heavy minerals of any kind. The improbability of the conjectured old Sewa
valley has thus once again been demonstrated.


The situation at the time of writing can be summarised as follows:-

1) Sound reasons have appeared fro questioning the conventional assumption that all Sierra Leone
alluvial diamonds originated from either the Koidu or the Tongo group of true kimberlites. The weight
of evidence collected during the present survey is considered strongly to favour the existence of
diamond sources in many parts of the fields.

2) It is established that kimberlitic indicator minerals are in general rare or absent, except in the
vicinity of the Koidu and Tongo kimberlites. If it is accepted that many local sources exist, then it
follows that they must be deficient in the normal indicator minerals. In one case at least, that of the
isolated Malema depsoits (Block 31), the data will admist of no other interprestation.

3) There is no real reason why the indicator-deficient source rock should resemble normal
kimbelrite. Indeed the converse is ture; if it did resemble kimberlite it would surely have been noticed
in the course of the Geological survey examinsations of mining sites.

4) Recent Geological Survey investigations have failed to discover an example of the indicator-
deficient source rock. Since its nature and appearance are completely unknown, more failures are to be
expected before the rock is finally identified.



(a) Prospects of success

The discovery and exploitation of kimberlite bodies is likely to become of increasing importance
in sierra Leone as the alluvial reserves diminish. Some small kimberlite pipes exist of coursw within
the S.L.s.T. Mining Leases and it may be that other, larger, ones will soon be found there. However,
the question to be examined here is whether there is any possibility of large pipes existing outside these
Lease areas.

In the established alluvial fields of south-east Sierra Leone, the comprehensive surveys already
carried out, with negative results, have shown with a far degree of certainty that no normal kimberlite
pipes exist in the licensed mining districts. However, the various isolated diamond discoveries made in
other parts of Sierra Leone show that most of the country is potentially diamondiferous terrain and
therefore merits explorlation for kimberlites, even though important deposits of alluvial diamonds may
be found only in the south-east. Experience in other parts of the world has shown that diamondiferous
kimberlite bodies are not necessarily associated with alluvial diamond fields. Alluvial fields are related
to kimberlites or other source rocks which have been deeply eroded, and whilst it is important to find
these sources, it is of even greater interest to locate kimberlite bodies which have suffered little erosion
since intrusion. In any diamondiferous terrain such as Sierra Leone there is a chance that these may
exist, particularly on the older planation surfaces. On a remnant of the pre-Eocene surface, for
example, even the original explosion crater of a Cretaceous kimberlite diatreme might still survivie in
the form of a large depression, such as that occupied by Lake Sonfon. Under such circumstances the
usual alluvial dispersion train of diamonds and indicator minerals could be negligible or absent.

It is considered that there is, in sierra Leone, a reasonable possibility that thorough exploration
will discover a major kimbelrite body.

(b) Recommended areas

In the exploration for kimberlite pipes, it is recommended that attention be directed principally
to the plateaux of the north-east quadrant of Sierra Leone i.e. the higher planation surfaces. This region
lies immediately to the north of the Koidu kimberlite field, where the geological environment is
identical, and drainage alignments reflect the widespread occurrence within the region of structural
features corresponding closely in direction with those associated with the kimberlites of Tongo and

In view of the apparent scarcity of alluvial diamonds in this region, it is unlikely that extensive
zones of kimberlite dykes crop out, and the exploration should be carried out on the assumption that the
targets are isolated pipes of kimberlite breccia.

(c) Exploration methods and costs

There is at present no single exploration technique which can be relied upon to find kimberlite
pipes in Sierra Leone. Combinations of methods must be devised to suit the circumstances. The
methods which can be employed are:-

1) Heavy mineral survey. This method consists of collecting heavy mineral samples, first from the
drainage and subseqnetly from soils, in order to detect concentratins of the kimberlitic indicator
minerals, pyrope and magnesian ilmenite. It has been successfully used to trace kimberlite dykes in the
Koidu and Tongo areas. Its limitations are:-

The indicator minerals are not always positively identifiable in the field.

It leads to barren kimberlites as well as to diamond-bearing ones.

Kimberlites which are deficient in the indicator minerals are not detected.

Kimberlites which have undergone very little mechanical erosion are unlikely to be detected.

It is apparent that a heavy mineral survey is, by itself, inadequate and should always be
supplementedby other exploration methods.

2) Analysis of diamond distribution. Diamond is the best kimberlite indicator mineral, as
only the last of the four limitations referred to above applies to it. It is considered that any kimberlite
search should include basic alluvial prospecting and that where diamonds have been found, the
distribution of values and the diamond characteristics should be closely investigated. The disadvantage
of the method is that alluvial concentrations are often related to ancient drainage patterns and may
therefore occur in all directions from the source, upslope or downslope on the present topography. For
this reason, the diamond distribution generally indicates a large ill-defined target area.
3) Examination of unexplained topographic depressions. This is of unproven validity in Sierra
Leone, as there was no identifiable topographical feature common either to all or to a majority of the
five small pipes found so far. It is thought, however, that the outcrop of a major kimberlite pipe among
granitic rocks must be indicated by a swampty depression, unless it is under a river channel or a flood
plain. A noticeable feature would certainly be formed by the traces of an explosion crater, but the
chances of these surviving are rather small. The identification of unusual topographical features is best
carried out on airphotos.

4) Geophysical methods. Ground magnetometer traverses over the Koidu kimberlites, carried
out by Sierra Leone Selection Trust in 1949, produced anomalies of up to 300 grammes. These
however, tended to be swamped by other anomalies produced by variations in the concentrations of
magnetite in the soils overlying the granite country rock. This effect will almost certainly be found
throughout Sierra Leone, and ground magnetometry is therefore not recommended. Aeromagnetic
survey is also of doubtful value, because where it has been used in connection with kimberlite surveys
in other territories the results have been either negative or inconclusive. Resistivity and gravity
methods should both be useful for the precise location of large kimberlite bodies whose position is
approximately known or suspected but whose outcrops are concealed by alluvium or laterite, but they
would appear to have little application during the exploration stage.

In view of the very high drainage density in Sierra Leone, an exploration programme which
combines systematic alluvial diamond prospecting with closely supervised heavy mineral sampling
should succeed in finding all normal kimberlites which have undergone some erosion. However, in
those cases where there has been little mechanical erosion (if they exist), there may be virtually no
alluvial dispersion of diamonds and other heavy minerals. An exploration programme must therefore
include provision for the identification and close examination of any topographical features which
might indicate the presence of pipe outcrops or crater traces would otherwise be missed.

The cost of this type of kimberlite exploration programme is estimated to be about Le.44,000 per
1000 square miles. This would cover alluvial diamond prospecting by the methods outlined in Part I,
Section E, and heavy mineral sampling at 1000 yard intervals throughout the drainage. It would not
cover follow-up work on discoveries.


(a) Suggested target areas.

Primary diamond deposits, or source rocks, which differ in many respects from normal
kimberlite are believed to exist in most parts of the established alluvial diamond fields. The evidence
on which this belief is based has been outlined in Section D. There is no part of the fields that should
not eventually be explored for these source rocks when their nature has been fully established, but the
areas listed below have been selected because they offer the best prospects of major discoveries. The
suggested initial target areas are:-

Block 10. Kagbwema area
Block 11. Ndogbogoma area. Kokoye valley. Segyei valley.
Block 13. Kponima area. Nyandeyama area.
Block 21. Foidu area.
Block 26. Folu valley, north of Putehun.

(b) Methods of search.

The first identification of the unknown source rock is expected to present some difficulty. It is
certainly deficient in the usual indicator minerals and may well be difference in appearance from
kimberlite; it could be an apparently unremarkable basic or ultrabasic rock representing a meta-
kimberlite. In a few cases (e.g. Kokoye, Segyei) there are small quantities of magnesian ilmenite
present which will assist the search, but generally speaking most of the techniques available for normal
kimberlite exploration will be inapplicable.
The only method which would appear to be of use is the analysis of alluvial diamond distribution and
characteristics, followed by bulk sampling of any unusual formation in the bedrock which appeared to
be related to a value cut-off or to a concentration of diamonds of a particular type. This will be slow
and expensive. It will initially be necessary to concentrate resources on finding one example, however
small, investigation of which may disclose some feature on which a more precise exploration method
can be based.




It is unfortunately true to say that from the very beginning of native alluvial diamond mining,
virtually no records of diamond recoveries or production from individual deposits have been kept. In
the early years, when all mining was illicit and all production was smuggled out, this was inevitable,
but even with the introduction of licensing it has been impracticable to make the miners keep
production records since most of them are illiterate and they are scattered through hhundreds of mining
areas. If Government were to attempt to supervise each mining site continuously and record production
on the spot it would need a staff of over 1000 trained men. Diamond dealers do keep record books,
from which the Mines Division abstracts data, but as many diamonds change hands two or three times
before reaching a licensed dealer it is only occasionally possible to trace a particular purchase back to a
mining site.

The only concrete data available on licensed alluvial diamond production in Sierra Leone are the
figures for purchases of diamonds by the Government Diamond Office, but none of these purchases can
be related to individual deposits or even to a particular mining field. Furthermore, total G.D.O.
purchases do not correspond with total alluvial production, since a proportion of this production is
smuggled out to the Monrovia market. In recent years, however, this proportion is thought to have
been very small. One of the principal objects of the present survey was the collection of data from
which estimates of production and remaining resources could be made.

The estimates of diamond resources which appear below and in the Appendix do not correspond
exactly with any of the reserve categories normally used for mineral deposits. Over 600 individual
deposits have been examined, and in about 400 of them one or two days have been spent on the
treatment of samples of miners’ gravel. It has not been possible to employ normal alluvial valuation
procedures as this would have entailed spending about one month on each deposit, or forty years on the
complete survey. There are consequently no Proved Reserves. All the estimates given can be regarded
as corresponding roughly with the category of Indicated Reserves as often used in formal reserve
calculations i.e. reserves for which values and yardages are computed partly from specific
measurements, widely-spaced samples and production data and partly from projection for a reasonable
distance on geological evidence.


(a) Production estimates

In order to estimate total production from a given deposit, it is necessary to find the gravel
yardage extracted, the values present in this gravel and the recovery rate of the miners. This is
comparatively straightforward where a deposit is being actively exploited for the first time at the time
of the geologist’s visit, but is rather more difficult where, to take the extreme but nevertheless common
case, the deposit has been mined several years before and is now flooded and covered by secondary
bush. In order to obtain the required data, combinations of the methods listed below have been
employed, depending on the conditions found. It will be appreciated that although the accuracy of the
data so obtained is in many cases doubtful they are still the best approximations available. The
following methods have been used:-
Sampling of miners’ gravel.

Pitting to obtain gravel samples.

Examination of licence-holders’ receipts for diamond sales.

Sampling of tailings.

Inspection and measurement and testing with the gravel probe.

(b) Reserve estimates.

Reserve estimation is concerned with the diamonds present in two types of ground:-

1. Gravel pillars, tailings, top bedrock and overburden in formerly-mined deposits.

2. Virgin alluvial ground.

A proportion of the diamonds are dispersed in ground which is unpayable by any standards, and for
these the term “diamond resources” is more appropriate.

In partically-mined ground, computation of the remaining reserves and resources has been
carried out concurrently with the estimation of production, and has been based on the same data,
however obtained. In virgin ground, average gravel thickness and values have been conservatively
extrapolated from mining sites or pits within the deposit or from an adjacent similar deposit, and only
the more reliable data have been used i.e. values obtained by sampling miners’ gravel or from
Geological Survey pitting.


All diamondiferous ground has been classified as either payble, marginal or unpayable. With
the exception of the “Payable M” category, the classification is based on the value of the deposits to
licensed alluvial miners and not to mining companies. The precise limitsof each classification vary
from one deposit to another, depending on local conditions, but the following generalizations may be

Payable ground. Gravel containing 0.4 carats per cubic yard is considered payable under any
circumstances. 0.3 carats is payable under medium overburden (6 feet to 15 feet), and 0.2 carats is
payable in shallow swamps or high terrace. By repeated mining of payable ground, a recovery of
approximately 70% of the total diamond content is eventually achieved by the licensed miners in most

Payable reserves have been subdivided into Payable A and Payable M ground. Payable A is
ground that can be tackled reasonably well by the licensed miners and will therefore eventually be
mined by them. Payable M is ground from which the gravel can only be economically extracted by a
fully mechanised operatin involving heavy capital expenditure and it is therefore suitable only for
exploitation by a mining company of substance.

Marginal ground. Deep ground with values between 0.3 and 0.15 carats per cubic yard, or shallow
ground with values between 0.2 and 0.08 carats is considered to be marginal. Such ground is
unpayable as a whole but because of the erratic distribution of diamonds a proportion of them can
always be recovered at a profit by the selective mining at which the licensed miners excel. The overall
recovery rate of diamonds in marginal ground is estimated at 40%. The production from such deposits
is one of the beneficial features of the alluvial mining scheme, since no mining company could produce
diamonds from such low-grade ground at a profit.

Unpayable Ground. This is ground with values below those of the marginal range as noted in the
paragraph above. Such ground is rarely mined and even if extracted during prospecting is usually left
untreated. Much of the unpayable ground consists of the mixed tailings, overburden and gravel pillars
in old mining areas; a little production is accasionally possible by the selective working of such ground
and it is estimated that about 5% of the diamonds in them will eventually be recovered.

Production and reserve estimates for each deposit (Table 7), Block summaries (Table 8), and
annual production statistics and smuggling estimates (Table 9) are all set out in the Appendix to this
Bulletin. The significant totals from these tables are as follows:-

Estimates Total alluvial diamond production, 1950 to 1965.

(excluding S.L.S.T.) – 10,537,000 carats

Total carats smuggled, 1950 to 1965 – 4,945,000 carats

Total remaining alluvial diamond resources at end
1965 (excluding S.L.S.T. Mining Leases) – 16,615,000 carats

Statistics Total diamond production of Sierra Leone Selection

Trust, 1932 to 1965 – 19,596,225 carats

Total alluvial diamond purchases by Diamond
Corporation and government diamond Office,
1956 to 1965 (excluding purchases from S.L.S.T.) – 7,880,739 carats


There is an obvious discrepancy between the total production from the alluvial diamond mining
areas, as estimated from the examination of the deposits, and the sum of legal diamond purchases and
illegal diamond exports. This discrepancy arises from the existence of a production source whose
magnitude it is not otherwise possible to estimate, that is, illicit mining and other forms of theft within
the S.L.S.T. Mining Leases. A reconciliation can therefore be made as follows:-

Total sales to Diamond Corporation (1956-1959) and
Government Diamond Offie (1960-1965) (Table 9)

Estimated illegal exports, 1950-1965) (Table 9)

– 7,880,739 carats.

– 4,945,000 carats


=12,825,739 carats

Total non-S.L.S.T. diamond sales

Estimated total production from alluvial mining
Fields, excluding S.L.S.T. areas (Table 8)

Assumed total output from illicit mining
Operations and theft within S.L.S.T. Mining Lease
Areas, 1950 to 1965

Total non-S.L.S.T. diamond production

– 10,537,100 carats

– 2,288,639 carats*

=12,825,739 carats

If the above estimate of total non-S.L.S.T. production is added to the factual of S.L.S.T.
production (Table 9), the grand total of Sierra Leone alluvial diamond production from all sources to
the end of 1965 is obtained, and this is 32,421,964 carats.

* This is a personal estimate by the author and cannot be regarded as an official one

Owing to the proportion of diamonds which are in unpayable ground, and varying recovery
rates, the net diamond reserve or potential production is much less than the total diamond resources.

The net reserve is estimated as follows:-









Export Value



Payable “A”




Payable “M”




















In the above estimates, the following assumptions have been made:-

Value per carat: Licensed miners’ production – Le.28.00 per carat.

Mining Company (mechanised) production – Le.24.00 per carat.

Revenue: Licensed miners’ production – Le.1.70 per carat.(Export duties, mining and

dealing licences).

Company (mechanised) production – Le.4.00 per carat. (Based on average annual

Revenue from S.L.S.T. production).

The principal alluvial deposits which still contain substantial diamond reserves and to which the
attention of interested parties is therefore drawn are:-



Estimated pay
Gravel yardage

(cubic yards)





Pools in Sewa channel.

Sewa channel, especially pool at S. end.

Flats and low terrace.

Sewa channel, flats and low terrace.

Few pools below elbow bend. Pool below Bohun rapids.

Few deep Sewa pools

Sewa flats and low terrace

Matemu flats

Sewa channel, many pools

Flats and low terrace

Bebeye/Sewa high terrace

Sewa channel.

Sewa flats and low terrace.

Sewa channel below Sumbuya

Sewa flats and low terrace.

Flats of Mamaye at Foindu.

Male channel and banks.













































For further details, reference should be made to the Block descriptions (Part III) and to the detailed

reserve estimates (Appendix).


(a) Production by Sierra Leone Selection Trust.

The Company’s production in 1965 was 652 thousand carats and this figure is also the average
for the last ten years*. The author’s estimate of the total remaining resources in the Lease areas
indicates that production from alluvial deposits will in fact continue for a period of between twelve and
fifteen years.

Production in the long-term depends on the underground exploitation of kimberlite bodies. Very
substantial underground resources exist, almost certainly exceeding the original alluvial resources, but
it is not yet established whether any of the know kimberlite bodies can be mined at a profit. If the
dykes and pipes are both found to be workable a steady underground production will lresult and the life
of the mines will be almost indefinitely extended.

(b) Production by licensed miners

Annual production by the licensed miners has been about 800 thousand carats for the last three
years.* Total potential production from the known remaining resources is estimated at 3,676,000
carats, which is sufficient to maintain production at present levels for about 4½ years if no new fields
are found. The probable pattern of production is a progressive decline over a period of ten years, after
which it might be expected to level off at about 50,000 carats per annum for a further ten years.

This projected decline is of little importance from the point of view of the loss of direct revenue,
which is in any case small (Le. 1.70 per carat), but it will cause distress among the alluvial miners,
about 20,000 of whom will lose their livelihood.

(c) Possible production from deep alluvials.

It has here been estimated that, within the established alluvial diamond mining fields, there are
6,683,000 carats of diamonds in deep but payable ground. These are the Payable “M” reserves and
they are suitable only for exploitation by mining companies. Most of them will require heavy capital
investment in dredges or other equipment, and as values in general are only moderate the profitability
of the mining operations will depend on a large gravel throughout and a high degree of efficiency. If a
means can be found of realising the potential of these deposits, the resulting production could be
between 200,000 and 300,000 carats per annum and this would do much to offset the ffects of declining
production by licensed miners.


(a) New alluvial fields for licensed mining.

It is unlikely that any major alluvial fields of the Sewa Valley or Tongo field type remain
undiscovered. It is most probable, however, that there are still many small fields to be found similar to
that of Malema in Block 31, and they may occur in areas far removed from the present mining fields.
Early opening-up of such fields is essential if the forecast decline in production by licensed miners is to
be averted. The miners themselves may make some new finds, but their chances of doing so unaided
are not very high and the wide-spread development of new fields is only likely to result from a
systematic alluvial diamond prospecting campaign undertaken by Government.

* This related to 1965
(b) Underground resources.

Kimberlite bodies have been show to existr within the S.L.S.T. Mining Leasre areas: exploratory
development of these underground diamond resources has now begun and it is expected that additional
important discoveries will be made in these areas during the next few years.

However, it has been contended in this Bulletin that diamond source rocks must also occur in
those parts of the diamond fields that fall outside the Lease areas. It has further been suggested (Part
IV, Section E (i) that the whole of the north-east quadrant of Sierra Leone merits exploration for
kimberlite pipes, even though alluvial diamonds may be scarce there. If these predicated kimberlites
and other source bodies do in fact exist in significant numbers then substantial potential sources of
revenue are lying undeveloped and unsought. It is clearly in the interests of the economy of Sierra
Leone that the necessary exploration should be expedited and encouraged.

Andrews-Jones, D.A. (1966), Geol. Survey of Sierra Leone, Bull. No. 6

Barber, M.J. (1961), “The Evaluation of a portion of the partially abandoned swamp, Sembehun 14”,

Sierra Leone State Development Company Limited, confidential lreport to Government.

Barber, M.J. (1963), “The Diamond Potential of some Previously Worked Swamps”, Diamond

Exploration Company (S.L.) Limited., confidential report to Government.

Dixey, F. (1919), “Pleistocene Movements in Sierra Leone”, Trans. Geol. Soc. South Africa, Vol.

XXII, pp. 112-117.

Fairbairn, W.C. (1965), “Licensed Diamond Mining in Sierra Leone”, Mining Mag., Vol. 112, No. 3,
pp. 166-177.

Forristal, C.J. (1965), “The Sewa Dredge Experiment”, diamond Exploration Company (S.L.)Ltd.,
confidential report to Government.

Geological Survey Department, Sierra Leone, annual Reports 1956 to 1965.

Grantham D.R. and Allen, J.B. (1960), “Kimberlite in Sierra Leone”, Geol. Survey of Sierra Leone,

Short Paper No. 8.

Haggard, H.J.E. (1958), „Explanation of Cross Section of Typicla Valley in W. African Humid Zone “,

Sierra Leone Selection Trust Ltd., notes for prospectors.

Haggard, H.J.E. (1963), “Review of the Yengema Diamondfield”, S.L.S.T. Ltd., unpublished internal

Company report.

Haggard, H.J.E. (1965), unpublished notes and personal communication.

Horton, R.E. (1945), “Erosional Development of Streams and their Drainage Basins”. Bull. Geol. Soc.

America, Vol. 56, pp 275-370.

Junner, N.R. (1946), “Preliminary Report on the Sierra Leone Diamond Fields”, S.L.S.T. Ltd.,
unpublished internal Company report.

Lovering, J.F. and Compston, W. (in preparation), a study of the radiometric ages of kimberlite

Marmo, V. (1962) Geol. Survey of Sierra Leone, Bull. No. 2

Michel, P. (1959), “L’Evolution Geomorphologique des Bassins du Senegal”, Revue de

Geomorphologie Dynamique, May-December 1959, vo. 10, pp 117-143

Mines Department, Sierra Leone, Annual Reports 1956 to 1965

Pollet, J.D. (1937), “The Diamond Deposits of Sierra Leone”, Bull. Imp. Inst., Vol 35, No. 3, pp 333-


Pollet J.D. (1952), “The Geology and Mineral Resources of Sierra L~eone”, Col. Geol. And Min.
resources, Vol 2, No. 1, p 3-28.

Stracke, K.J. (1963), “The Prospecting for Diamondiferous Kimberlite in Sierra Leone”, Diamond

Exploration Company (S.L.) Ltd., confidential report to Government.
Strahler, A.N. (1952), “Hypsometric Analysis of Erosional Topography”, Bull. Geol. Soc. America,

Vol. 63, pp 1117-1142.

Van der Laan, H.L. (1965), “The Sierra Leone Diamonds”, Oxford University Press.

Wells, M.K. (1962), Geol. Survey of Sierra Leone, Short Paper No. 9.

Wilson, N.W. and Marmo, V. (1958), Geol. Survey of Sierra Leone, Bull. No. 1

Wilson, N.W. (1965), Geol. Survey of Sierra Leone, Bull. No. 4