·    Four types of drilling used: hand auger, vibracore, air core RC, and sonic drilling. 85% of the drilling is aircore RC.

·    Drill samples taken either at 1.5m or 3m intervals.

·    Total Heavy Mineral (THM), 1mm oversize and -0.045mm "slimes" fractions determined on all drill samples. 

·    Mineralogy of the THM from selected drill composites determined by QEMscan and XRF.

·    All hand auger samples were collected over 1.5m intervals with depths rarely exceeding 15m. After retrieval, samples were placed in calico or canvas bags and labelled with the hole number and sample interval.  An inherent problem with the hand auger technique is over sampling, with the collected sample interval presenting a larger volume than the theoretical volume. Hand auger drilling is also impossible in clays or wet sands below the water table.

·    Reconnaissance hand-auger drilling is mostly used to locate the major anomalous mineralised areas. Sample quality is low compared to other drilling methods and hand-auger drillholes have been superceded with later drilling of better quality. However the hand-auger drillholes contain valuable mineralogy results, so they have been retained for the final block model.

·    85% of the drilling consists of reverse circulation air-core drilling with NQ hole diameter approx 75mm. Four different RC rigs were used. The size and capabilities of the different RC drill rigs change with the air compressor capacities between 1000kPa and 1600kPa. The RC rigs operate using compressed air as the flushing medium which is sent down the borehole between the outer and inner tubes. Although water can be added to aid sample recovery no polymer was used during drilling. Tungsten carbide tipped bits are used to advance the face of the hole and all holes were drilled vertically.

·    Reverse circulation drill samples were collected over 3m intervals. A pocket PC was used with the Fieldworker software, where all samples were described and field estimates introduced. Sample logs provide information such as description of sediments (colour, grain size, sorting), field estimations for THM, interval, depth and comments on variations of the sedimentary and geomorphologic characteristics of the terrain.

·    Vibracore drilling was used on wetlands or areas where the water table is shallow (6% of the total metres drilled within the resource boundary). The Vibracore unit comprises a 1m long sampler with OD 50mm and ID 44.4mm and 1.5m length rods. The technique operates by applying a vibrating motion to the surrounding sand through the application of an oscillating motion to the drill string generated at the drill head using a motor. The vibration essentially liquefies the ground and the drill penetrates under its own weight without the need of rotation.

·    Sonic drilling employs a resonance-vibratory technique with high frequency mechanical vibration and no flushing medium to take continuous samples and to advance drilling rods into the ground.

·    The Sonic rig employed in Mozambique used a '4 x 6' system in which a 4 inch (100mm internal diameter) core barrel is advanced 1.5, 3 or 6m followed by a 6 inch casing string which is advanced over the 4 inch core barrel to the bottom of the hole. Commonly 6m or 6m runs were used with the choice of core barrel  advancement dependant on recovery i.e. 6m run used when experiencing good recovery, and a 3m run used when recovery was poor.

·    Sample quality from Sonic drilling is excellent and the objective of the sonic drilling programme was to obtain samples that would test the validity of the RC, Vibracore, and Hand Auger samples, in addition to verifying the accuracy of the different drill rigs and contractors that completed earlier drill programmes.

·    Hand auger samples are commonly contaminated with material falling down the hole as drilling proceeds. Samples that were larger than expected were logged as possibly contaminated.

·    In Reverse Circulation drilling the sample is retrieved by air flushing where the sample is blown up the inner core barrel through a collection hose into a plastic sample bag. The samples were collected in heavy duty plastic bags held in a cyclone close to the drill rig (different types of cyclones were designed for different drilling programmes). The whole sample was collected in the field.

·    Dry sample weights were recorded at the laboratory and are a measure of sample recovery. The theoretical weight of a 1m NQ sample is about 7kg. With this drilling method, there is normally lower than average recovery at the top of the drillhole due to sample losses into the surrounding soil (the air pressure of the system, combined with the weight of the rods and relatively low back-pressure of the surrounding sands causes sand near the bit face to be pushed outwards into the surrounding ground). Additionally, below the water table sample recovery can be greater than 100% as water flow causes the hole to have a greater diameter than the drilling bit. Good drilling techniques are used to minimise the effects of both of these problems. During RC drilling, the best samples are "moist lumps" where there has been no possibility of fines segregation (blow-over) or losses of THM due to separation in water.

·    For the Vibracore drilling the drill string is advanced in 1.5m runs with the sample retrieved from each run using a hand pump. The hand pump has a gravity valve which is designed to allow sample to be pushed into the sampler and prevents it from being lost. The sample retrieved by the sampler is placed in buckets until the full depth of the interval is reached and then transferred to sample bags. Samples were dried and weighed at the laboratory, giving a measure of sample recovery.

·    Samples are retrieved from the Sonic drilling using a core barrel. Samples are vibrated out of the core barrel into a plastic sleeve, and present very much like a core sample (see the picture to the left). There is normally a small amount of sample compression with the sonic technique (associated with the wall thickness of the coring tubes), and the length of core retrieved is sometimes slightly longer than the penetration depth. Sonic drilling provides the best sample recovery and sample quality of all of the methods.

·    Hand auger samples were recorded on paper field logs or electronically with Pocket PC devices.

·    RC samples were logged on-site using either a paper log or Pocket PC using the Fieldworker program. Sample logs provide information such as description of sediments (color, grain size, sorting), field estimates of THM, interval depth and comments on variations of the sedimentary and geomorphologic characteristics of the terrain.

·    Logging of the Vibracore samples was the same as for the reverse circulation and hand auger holes, all samples were logged on-site using a paper log or Pocket PC.

·    At the Sonic drilling sites, only small pinch was taken from the bottom of the core for description of the sample, estimation of THM% and deciding if the hole should be terminated. For each 3m or 6m interval of sample, this information was transcribed to a paper log for quick reference prior to the completion of the logging at camp.

·    Detailed sample logging was then completed at camp. The samples collected were logged in 3m intervals. For 6m runs, the measured length of the run was split in half producing two separate 3m samples. Each sample was logged using a paper log sheet which was later entered into the field database.

·    Hand auger samples were homogenised and then reduced using "cone and quartering" splitting technique in the field.

·    For the RC, Vibracore and Sonic drilling methods, the whole drill samples were delivered to the laboratory.

·    After delivery to the Inhambane laboratory, samples were checked, dried and then riffle split down to a nominal size of 350g.

·    This split size is high (ie more conservative) relative to other operators globally, where heavy mineral sand sample splits of 100 to 200g are more common.

·    Sample duplicates were inserted at rate of 1:15 to check laboratory accuracy, with good results.

·    Sonic drillholes were used to "twin" earlier drilling as a check of in-situ values. The results showed good repeatability for the RC drilling, but only fair results for the hand-auger drilling. This is expected given the limitations of the method.

·    Sieving to determine 2mm (oversize) and -45micron (slimes).

·    Heavy mineral separation using heavy liquid to separate THM from other minerals (predominantly quartz).

·    Control procedures include laboratory duplicates, blind duplicates and standard samples. 

·    In Rio Tinto labs, LST was used and density is monitored and kept above 2.85 Samples prepared at SGS were separated in TBE diluted with acetone to maintain a 2.85 density.

·    64 twinned holes completed. Sonic drillholes used to check other types of drilling show good correlation with RC and Vibracore drilling, but lower correlation with hand-auger results (this may reflect the lower survey accuracy of the hand auger holes than problems with the drilling technique itself). Twin holes of RC/RC and Vibracore/Vibracore types show good correlation.

·    Most of the field data is entered into computers in the field. In places where paper logs are used, the information has been transcribed and entered into databases.

·    Samples are delivered to the laboratory with a sample submission form.

·    At the laboratory, samples are checked against the submission document and entered into the laboratory systems.

·    Different laboratories were used. Those with the best practices used barcode control and scales linked to the computer systems (laboratory information management systems - LIMS). Other laboratories recorded results by hand and transcribed them into the database systems. These systems are inherently less reliable due to typographical and hand-writing errors, however at the time of the analyses, this was the method used at many major commercial laboratories.

·   Assays and other drilling data are recorded in the Acquire data management system, which includes many integrity checks.

·    No adjustments have been made to the laboratory data.

·    The majority of holes used in the model were surveyed using differential GPS. Hand-held GPS was used to locate drillholes in the early phases of reconnaissance and exploration. Levels for drillholes not accurately surveyed were determined using LIDAR.

·    Drillholes are positioned using UTM zone 36S, WGS84 datum.

·    Despite the highly variable dune topography across much of the region, excellent control has been established using the LIDAR survey.

·    Drill spacing for much of the resource is 250 x 250m, although zones along the eastern side of the resource are at 350 x 350m drill spacing and a small area in the north-eastern section of the resource is drilled at 500 x 500m. Variography shows that the bulk of the resource lies within the variogram range of at least one drill sample. Areas that required a larger ellipsoid are found along the very edge of the resource, and in the 500 x 500m drilled zone in the north-east.

·    Sample composites were used to determine mineralogy of the THM.

·    Three geological units host the bulk of mineralisation: Dune 2, Dune 3 and Fluvial Unit.

·    Dune 2 shows no preferred horizontal direction in mineralisation trend.

·    Dune 3 and the Fluvial unit show a north-south trend. The drilling pattern does not preferentially sample the mineralisation in any unit.

·    The largest risks to the samples are weathering of the containing bags and damage or loss to the bags during transport.

·    Large resistant plastic bags were used for the vibracore and reverse circulation samples. They were locally supplied and the quality was not always the same. On several occasions they were doubled to avoid mass loss during sampling. The sample identification number was written on aluminium tags placed inside the bags, in plastic tags with cable ties closing the bags and on the external part of the plastic bag. Samples collected at drill site were transported in the same day to camp sites, and within just few weeks transported to the laboratory warehouse. In camp site they were exposed to sunlight and rain, but were protected from the weather in the laboratory warehouse. During transport, a chain of custody was followed with sample dispatch forms, and  samples received were reconciled with samples shipped. Nevertheless, handling and loading/unloading sample bags onto open trucks and pick ups sometimes caused damage to the bags. This damage resulted in loss of material and sample refusal by the laboratory. Improvements of the transportation were only achieved with constant monitoring by the supervisors and foremen.

·    Exploration Licence 1336L, of area 118.8 km2, lies approximately 25km south of the regional capital Inhambane and approximately 340km north east of the national capital Maputo.  The lease is held by Rio Tinto Mining and Exploration Ltd, and subject to the joint venture agreement with Savannah Resources. The lease is due to expire on the 21 January 2018.

·    There are no known impediments to mining development, other than the normal social issues regarding relocation - if necessary.

·    Rio Tinto have conducted multiple phases of exploration on the area since 2000. The area was previously known to be mineralised.

The Inhambane region contains vast quantities of reworked coastal sands that were deposited by the Limpopo River further south. Mineralisation at Jangamo is hosted in dune sands 6 to 10 km inland from the current coastline. Within the Jangamo area, four major units are recognised: Dune 1, Dune 2, Dune 3 and a Fluvial Unit. Most of the mineralisation at Jangamo is hosted by Dune 3 and the Fluvial Unit.

·    See Appendix 1 - drillhole intercept listing from the resource zone.

·    The grades of the drillhole intercepts listed in Appendix 1 are the average grades for the drillhole intercept of the resource.  The resource is defined using the JANC block model with a boundary determined by maximising revenue from the model, assuming a set of costs, recoveries and revenues for the contained ilmenite, rutile and zircon (assumptions are listed below). The boundary generally matches the 1.7% VHM supported grade boundary. VHM is the sum of in-situ ilmenite percent, zircon percent and rutile percent.

·    The drillholes are vertical and the mineralisation is generally sub-horizontal. 

·    See Figure 1, Plan view Figure 2 Drillhole Plan and Figure 3, Jangamo THM Grade  Figures 4 - 10 representative sections.

·    All drillhole intercepts are listed in Appendix 1.

·    Airborne geophysics was used to help target drilling. Magnetic and radiometric data are useful to detect mineralised areas.  Detailed SRTM elevation data and LIDAR elevation data are also used to help interpret paleo-landforms.

·    Mineralogy was determined using QEMScan scanning electron microscopy on composite samples from the mineralised areas.

·    Slimes in the resource (the "fines" fraction, of less than 0.045 mm size) averages 7%. Smectites are only present in the basement.

·    No significant hard layers have been intersected in the drilling within the resource. Some paleosols between the dune units have elevated slimes.

·    Groundwater has been measured around the resource and a groundwater model created. In some areas the level of groundwater would need to be managed if it were being dry mined.

·    No "potentially acid forming" soils have been documented in the main mineralised zone. There is potential for some acid-forming materials to be present in the basement or small pockets of the fluvial unit.

·    Monazite levels are low.