Bulk Sampling

The objective of the bulk sampling at Tongo Field was to simulate as closely as possible the results that could be achieved during underground mining, i.e. sampling of hard rock kimberlite rather than decomposed kimberlite, and processing the whole sampling/mining width slot, including waste, in order to obtain a representative mining grade for each dyke zone. The narrowness of the dyke lenses (between 10cm and 37cm) within the 1-1.5m mining width means that significant volumes of waste (up to 90%) had to be mined and processed.

The diamonds recovered were cleaned with hydrofluoric acid, resulting in between 0.85% and 1.68% reduction in carats, with the highest loss recorded for the Kundu sample which contained fresh olivine crystals mistaken for coated diamonds and which were dissolved.


Artisanal diggings on the Lando Dyke Zone site	SLST excavation at Mavehun on the Lando Dyke Zone	Artisanal diggings on the Kundu Dyke Zone site


Excavation of the Lando Dyke Zone at Mavehun	View of the Kundu excavation during the overburden stripping phase	Lando Dyke Zone - kimberlite stringers cutting across fabric of gneissic country rocks

The following photos show aspects of the morphology of dyke zones and macroscopic views of the nature of the kimberlite.


Lando A mapping surface	Lando B mapping surface	Lando A dyke showing flow banding of olivine macrocrysts	Lando B stringers


Kundu mapping surface looking west	Charging of blast holes along the Kundu Dyke Zone	Kundu dyke showing aphanitic and coarse macrocrystic kimberlite	Close-up view of Kundu dyke showing aphanitic and coarse macrocrystic kimberlite

Sampling Method and Approach

Due to the constraints imposed on operations by the rainy season (May to October) and the large size of the licence area, Koidu Holdings elected to focus first on sampling the Lando and the Kundu dyke zones, historically reported to have the highest grades. One of the objectives was to verify the historical data and, hence, samples were situated along strike of the previous excavations.

The QuickBird satellite image was used to identify suitable locations for the bulk samples, which were or could be made accessible by earthmoving equipment and which would not significantly impact on the day to day living of the nearby community. The coordinates of the target areas were identified on the ground using a GPS, and the traces of the dykes could easily be recognised by the trenches that had been dug by the artisanal miners. Two sample sites were located on the Lando dyke zone, one at the old SLST plant and the other at Mavehun, where bulk sampling had been undertaken by SLST in 1964. The rationale behind the selection of the Mavehun site was to attempt to duplicate the results obtained by SLST at the same location. In addition, the historical maps of Mavehun showed coalescing of two dyke lenses, forming a small blow which appeared to reduce in size with depth.

Considering the extensive artisanal mining in the area and the degree of disturbance of the ground in the vicinity of the kimberlite dykes, the decomposed kimberlite and granitic country rock was removed so that samples of fresh, un-weathered, in situ kimberlite could be extracted. The weathered zone at the Lando excavation sites ranges from 7m to 14m, while at Kundu the weathered zone averaged 12m in depth.

The overburden stripping at the Lando SLST plant site was undertaken from 14 November 2005 to 28 February 2006. A total of 164,378 tonnes of waste and 611 tonnes of decomposed kimberlite was removed. While overburden removal was in progress at the SLST plant site, construction of an access road to the Kundu sample site commenced.

After dewatering of the flooded Mavehun excavation at the end of March 2006, excavation of overburden got underway and was completed by the end of April 2006. A total of 97,539 tonnes of waste and decomposed kimberlite was removed. As suggested on the historical maps, the Mavehun blow had reduced to narrow dykes at depth and further work on this area was not pursued due to the close proximity of houses.

Work on the Kundu overburden stripping commenced at the end of March 2006 and was completed by the end of May 2006.

The total volumes and tonnages of overburden removed from the Lando and Kundu excavations are tabulated below:

Stripping of Overburden

Place	Waste Tonnes (t)	Decomposed kimberlite (t)	Total Tonnes (t)
Lando (at SLST plant)	178,806	611	179,417
Lando (at Mavehun)	88,352	1,671	90,023
Kundu	97,539	-	97,539
Total	364,697	2,282	366,979

Once the overburden had been removed and the hard rock exposed, the surface was thoroughly cleaned so that the surface could be surveyed and dyke lenses and stringers could be geologically mapped and the contacts surveyed.

Of the 200m strike length planned to be extracted at Kundu, an end-to-end excavation measurement of 195m was achieved, while of the 100m each strike lengths of Lando A and Lando B, 97.9m and 92.4m, respectively, were finally excavated.

MSA Geoservices (Pty) Ltd, a South African based consulting company specialising in exploration in Africa, was contracted to independently supervise the bulk sampling programme and to ensure the representivity and integrity of the samples from the point of extraction to the conclusion of the sale. An MSA field geologist, Dave Rolfe, was stationed on site at Tongo Field from April 2006 to May 2007 and was responsible for:

Mapping the dyke exposure at the surface of the approximately 1m wide hard rock excavation, as well as at the base after sample extraction;
Photographing the dyke exposures;
Controlling the loading and hauling of the bulk samples to the stockpile at Tongo;
Controlling the loading and hauling of bulk samples from the Tongo stockpile to the stockpile at Koidu.

MSA's Project Manager, Stuart Hill, visited Tongo from time to time to review progress during the extraction phase and was present during the final sample processing at Koidu, to audit the plant and to monitor the chain of custody of the diamonds from recovery to sale.

Geological Mapping

The dyke zones were mapped by MSA using the conventional mapping techniques of establishing control points and recording the widths of the dykes and stringers on graph paper at 1m intervals. Koidu Holdings' surveyor then picked up points along the contacts of the individual dyke lenses, importing the points into Autocad where the contacts could be geo-referenced and traced accurately in 3D. The detailed survey of the dykes provided the basis for the kimberlite volume calculations.

The dyke zones were mapped in detail at the surface of the hard rock excavation, as well as at the base after the sample had been extracted, in order to detect changes in dyke width in the vertical and horizontal planes.

The excavation along the Lando Dyke Zone revealed two discrete dyke units, named Lando A and Lando B, each of which shows considerably more horse tailing and variability in dyke widths than the single main Kundu dyke.

Strike orientations of the Lando A dyke lenses range from 054° to 062°, with an end to end strike average of 056°. Lando B strike measurements range from 065° to 072°, with an average orientation of 066°. There is less variability in the strike orientations of the Kundu lenses, which range from 061° to 063°.

It is evident from the mapping that over a vertical distance of approximately 5m there is considerable variation in the morphology of the individual dyke lenses in the Lando A and Lando B dyke zones. The dip of Lando A changes from vertical in the south west to between 82° SW and 85° SW in the north east of the mapped exposure.

Lando B has a dip of between 80° SW and 85° SW, which is also more pronounced towards the northeast.

The dykes occupy regional fracture zones, with the kimberlite dyke contacts usually parallel to the fracture planes. Shearing along the contacts is preserved in places, where slidensides are visible. Calcite veining is variably developed parallel to the contacts, particularly at Kundu. The dykes and fractures cut across the regional fabric of the granite-gneiss country rocks.

The widths of the dyke lenses and stringers were measured by MSA at 1m intervals along strike at the surface and at the base of the excavations.

Satellite Imagery

The QuickBird satellite developed by DigitalGlobe offers highly accurate, commercial high-resolution imagery of Earth and is the only spacecraft able to offer sub-meter resolution imagery, industry-leading geolocational accuracy, large on-board data storage, and an imaging footprint 2 to 10 times larger than any other commercial high-resolution satellite.

The satellite image of the Tongo Diamond Field Project area was ordered in July 2004, but due to the overcast conditions during the rainy season, a clear view with less than 5% cloud cover was only acquired by the satellite on 8 February 2005. The image documents the extent of artisanal workings prior to commencement of activities, as well as the location and distribution of all dwellings within the licence area. The ArcView GIS software programme was used to view and interpret geological and geographic features, and was valuable during the planning stages of the exploration programme.

The image is available on the GoogleEarth website.

Diamond Valuation

The diamonds were transported to the Government Gold and Diamond Office in Freetown under independent security and accompanied by Stuart Hill of MSA for valuation. The standard procedure for valuation of Koidu parcels was followed. Each size fraction was valued separately by each of the three valuators, i.e. the GGDO valuator, the Steinmetz valuator and the independent government valuator Luke Marriot of DCI. The parcel was already sorted according to the size fractions required by the valuators. Each size fraction was re-weighed using the GGDO scale to confirm the declared carats. The average of the values for the Lando A was $163.14/ct, Lando B was $192.09/ct and Kundu was $202.74/ct

Application for a Mining Lease

A letter to the Director of the Geological Survey, dated 9 May 2007, provided the preliminary results of the bulk sampling and requested an extension of the exploration licence for a further one year period, to allow the Company to complete geophysical surveys, drill the targets generated and commence feasibility studies. Although the Mines and Minerals Decree allows for a three year period during which to complete the feasibility studies, the internal pre-feasibility studies already completed by Koidu Holdings indicate that mining the Lando and Kundu dyke zones would be economically viable, despite the challenges posed due to their narrow widths. For this reason, the Company intends to apply for a mining lease before the end of the permitted three year period. Koidu Holdings has shown itself to have the technical and financial capacity to develop mining operations and successfully complete bulk sampling programmes in Sierra Leone, and wishes to begin the Tongo Diamond Field Project development programme as soon as possible.