MEMORANDUM

Date: 30 June 2005

Head Office:
Company: Crescent Gold Limited 1162 Hay Street
West Perth WA 6005
Attention: Andrew Spinks P O Box 1671
West Perth WA 6872
Copy: Telephone: +61 8 9324 8800
Facsimile: +61 8 9324 8877
From: Neil Inwood Email: [email protected]
Web: http://www.rsgglobal.com
Subject: Lord Byron Scoping Estimate

As requested by Mr Andrew Spinks (Manager – Geology, Crescent Gold), the geology and gold
grade distribution were modelled for the Lord Byron deposit. The mineralisation was modelled using
a 0.3g/t Au cut-off, with a view to establishing a resource for a heap-leach mining operation.

The Lord Byron deposit is located 63km east of the Sunrise Dam Mine site, near Leonora, Western
Australia. The geology of the Lord Byron area is discussed in Payne (2001), and Kent (2004).

The following is a summary of the work undertaken.

DATABASE
The database used for the Lord Byron and Fish areas was supplied by Crescent, and validated
using standard RSG validation techniques. The resulting database was called RSGLordByron.mdb.

A previous database for the Lord Byron area was located which had been used for the 2001
resource estimate by Resource Evaluations (Reseval) (Payne, 2001). The two databases were cross
checked for critical records such as collar location, sample ID’s, assay values for samples, and
survey records. No significant errors were identified between the two databases.

Two RAB holes (JSPR0004 and JSPR0005), which fell within the mineralised area, were not used
for the grade calculation.

Issues initially identified with the supplied Lord Byron/Fish database included:

ƒ Duplicate hole ID’s with different ‘DataSet’ designations, indicating changes made by AGAA –
superseded holes where deleted from the database ;
ƒ Missing Hole Type designation - fixed by cross-correlation to the 2001 Reseval database;
ƒ Missing assays for hole JSPC1229 -obtained from the Reseval database; and
ƒ 2 holes had a collar R.L. of 0 or 500, - these were changed to suit the surface topography.
ƒ The supplied database only had gold values in the ‘AuPref’ field, with no indication as to their
origin (e.g. Au1, average). Comparison to the Reseval database indicated that the values were
the Au1 (i.e. first assays) result from the relevant laboratories.

P:\Projects\Crescent_Gold\PCRE10_Lord Byron_ResEst\Surpac resource june 2005\Reports\pm_PCRE10_ResourceMemo-Inferred_050711_01.doc Page 1 of 6

Due to incomplete data, additional QA/QC analysis to that performed by Kent (2004) was not
conducted. Analysis of QA/QC data will be required to improve the resource classification.

MODELLING
The mineralised portions of the Lord Byron geology were modelled based upon a nominal 0.3g/t gold
threshold, with a nominal maximum of 5m of internal waste. Surpac Vision was used for the
modelling. Mineralised zone continuity was assessed against lithology and observed grade trends.
Reference was made back to the modelling performed in 2001 by Anglo Gold Ashanti Australia
(AGAA) (Kent, 2004), and Reseval (Payne, 2001).

A total of 23 separate units were modelled. These were grouped into 5 domains for the purposes of
statistical and variographic analysis. The 5 domains were the Bicentennial Shear Zone (BSZ), BIF
units, laterite units, NNE trending shears, and supergene units.

The base of complete oxidation (BOCO) and top of fresh rock (TOFR) were modelled based upon
regolith logging codes in the supplied database.

GRID TRANSFORMATION
As the main mineralisation (BSZ) had a strike of 320 degrees, a rotated grid was created. All data
was transformed to the local grid. The grid transformation parameters are shown below.

Table 1
Local grid transformation for Lord Byron

AMG E AMG N AMG RL LB E LB N LB RL

504175.055 6776324.773 0 5000 10000 0
503885.338 6778692.486 0 6300 12000 0
Rotation: 40º
Scale: 0.999999931

DENSITY
Two versions of density data were coded into the database. These have been based upon values
used by Payne (2001) (Table 2), and Kent (2004) (Table 3). At this stage, the Payne (2001) density
values were considered appropriate to be used for the resource estimate. RSG Global suggests
further work is required to adequately establish robust densities for the Lord Byron area. Due to
incomplete density data additional density information will be required to upgrade the resource
classification.

Table 2
Payne (2001) Density Values

Material Shear Zone BIF

Laterite 2 2
Oxide 1.8 2
Transitional 2.2 2.2
Primary 2.7 2.7

Page 2 of 6
 Table 3
Kent (2004) Density Values

Material BSZ Laterite BIF UM Dykes Other

Saprolite 2.55 2.55 2.65 2.7 2.7 2.6
Saprock 2.7 2.7 2.85 2.8 2.85 2.85
Fresh 2.85 2.85 2.95 2.95 3.1 3.05

MODELLING PARAMETERS
The mineralised zones were coded into the geological database. The drill hole intersections for each
zone were composited to a 3m down-hole length, missing assay values were given a grade of
0.005g/t Au.

The univariate statistics for each domain were examined to investigate the population distribution,
and determine any high-grade cuts. A high grade cut of 6g/t Au was made to the BIF grade
population.

Directional variograms were calculated on the 3m composites for each domain using Isatis software
(Table 4). These variograms were used to calculate the relevant anisotropy for each domain.

Table 4
Variography Parameters

Unit BSZ BIF Laterite NNE Shears Supergene

Azimuth 0 40 30 0 0
Dip 40 0 0 0 0
Plunge -75 -75 0 0 0
Nugget 0.4 0.4 0.15 0.5 0.4
Structure 1
Sill 0.4 0.4 0.55 0.3 0.4
Range 20 45 17 20 17
Semi-Major Range 40 20 17 20 17
Minor Range 15 25 17 20 17
Structure 2
Sill 0.2 0.2 0.3 0.2 0.2
Range 110 55 70 30 38
Semi-Major Range 60 20 70 30 38
Minor Range 30 35 70 30 38

A block model (lordbyron20050628.mdl) was created using Surpac Vision on the Lord Byron local
grid. Parent cell dimensions of 20mN x 10mE x 5mRL were used. Sub-blocks down to 5mN x 2.5mE
x 1.25mRl were created so as to provide effective volume representation of all horizons. The
dimensions and attributes of the block model are summarised in Tables 5 and 6.

Page 3 of 6
 Table 5
Block model dimensions and sizes

Type Y X Z
Minimum Coordinates 10400 5300 0
Maximum Coordinates 11700 6000 500
User Block Size 20 10 5
Min. Block Size 5 2.5 1.25
Rotation 0 0 0

Table 6
Block model attributes

Attribute Name Type Decimals Background Description

av_dist Real 2 0 Average distance to samples
class_code Integer - 0 Confidence class
krig_au_cut Real 2 -99 au_cut value
krig_var Real - 0 Kriging variance
lithology Character - Rock lithology
min_body Integer - 0 Mineralised unit code
near_samp Real - 0 Distance to nearest sample
no_samp Integer - 0 Number of samples
pass Integer - 0 Estimation pass
rock_type_int Integer - -99 0=air, 1=waste, 2-mineralised
Densitd Real 2 -99 Bulk Density AAGA
density_rsg Real 2 -99 Bulk Density - Payne 2001
weather_type_int Integer - -99 0=air, 1=oxide, 2-trans, 3-fresh

GRADE ESTIMATION
Grade estimation was performed using Surpac Vision. Gold grades were estimated using the
Ordinary Kriging method. Optimal parameters for the estimation passes were based upon horizon
geometry, drill hole layout, analysis of the slope of regression, and analysis of Kriging weights.

At this at stage – scoping level – no change of support was applied for a particular selective mining
unit. RSG Global understands that Crescent expects to use a fairly large SMU for the heap leach
operation, possibly of the order of 20 by 10 by 5m, but that the final SMU will depend on the
outcome of mining equipment / selection sensitivity studies. For the current estimate, gold grades
were directly estimated in panels of 20 by 10 by 5m, which is roughly half the drill hole spacing,
thereby evading the need for change of support. If however, Crescent opts for a smaller SMU
without further infill exploration drilling, change of support will have to be considered.

The grade estimations were obtained from three passes. The first pass was designed to maximise
the slope of regression and minimise negative Kriging weights. The second pass was expanded in
size, with a reduction of required samples. The third pass was expanded in size to assign a grade
estimate to the remaining blocks. The pass parameters are shown in Table 7.

Page 4 of 6
 Table 7
Search Strategy

Unit BSZ BIF Laterite NNE Shears Supergene

Pass 1 2 3 1 2 3 1 2 1 2 1 2 3
Major Axis 80 120 200 55 80 120 60 80 30 50 38 38 60
Semi-Major Axis 53 80 200 28 40 120 60 80 30 50 38 38 60
Minor Axis 27 40 200 37 53 120 60 80 30 50 38 38 60
Minimum No. Samples 16 12 3 16 12 1 10 1 16 1 16 12 3
Maximum No. Samples 32 32 32 32 32 32 32 32 32 32 32 32 32

GRADE – TONNAGE ESTIMATE

The Inferred resource estimate, above a 0.5g/t cut-off, above the 350mRL is summarised in Table 8.

Table 8
Inferred Resource Estimate
Density values as per Payne (2001), at > 0.5g/t Au cut off, and above the 350mRL

Category Tonnes g/t Au Oz. Category Criteria

1 2,287,000 0.98 72,100 >350m RL, All units
2 1,838,000 0.98 57,900 >380mRl, Pass =1, BIF, Lat, BSZ
Total 4,125,000 0.98 130,000

RSG Global feels that pending validation of QA/QC data and verification of density data, the
category 2 mineralisation has the potential to be classified as Indicated resource.

COMPARISON TO PREVIOUS STUDIES

Table 9 summarises previous resource estimates performed on the Lord Byron deposit.

The Inferred resource estimated by RSG Global is higher than those previously reported. However
the Reseval 2001 total inventory of mineralisation above the 350mRL ,3.3Mt @ 1.3g/t uncut for
139,000 oz, is comparable to that obtained by RSG Global. The difference can be attributed to a
lower grade cut-off being used for modelling, 0.3g/t Au versus 0.5g/t Au, and the application of a top
cut.

Table 9
Previous Resource Estimates

Author Tonnes g/t Au Oz. Comments

Bateman (1998) (SOG) 2,256,000 2.2 160,000 Indicated and Inferred
Payne (2001) (Reseval) 1,673,000 0.8 41,600 Inferred and Strategic
Payne (2001) (Reseval) 808,000 0.8 20,783 Inferred only, >3g/t, >380mRL
Shaw (2004) ( WMC) 710,000 2.01 45,870 Inferred - 1g/t cut off, $650 whittle shell
Kent (2004) (AGAA) 434,751 1.43 19,991 Inferred - 1g/t cut-off, $650 Whittle shell

Page 5 of 6
REFERENCES

Bateman, J. 1998 in Payne (2001).

Kent, M. 2004. AngloGold Ashanti Australia Ltd Jasper Hills Project Lord Byron and Fish Resource
Estimates December 2004. Internal Company Report. AngloGold Ashanti Australia Ltd

Payne, P. 2001. Mineral Resource Estimate For The Lord Byron Deposit, Jasper Hills Project,
Western Australia. Internal Company Report. Resource Evaluations Pty Ltd.

With best regards

Neil Inwood
Senior Resource Consultant

This document has been prepared for the exclusive use of Crescent Gold Limited on the basis of instructions, information and data
supplied by them. No warranty or guarantee, whether express or implied, is made by RSG Global with respect to the completeness
or accuracy of any aspect of this document and no party, other than the Client, is authorised to or should place any reliance
whatsoever on the whole or any part or parts of the document. RSG Global does not undertake or accept any responsibility or
liability in any way whatsoever to any person or entity in respect of the whole or any part or parts of this document, or any errors in
or omissions from it, whether arising from negligence or any other basis in law whatsoever.

Page 6 of 6