PERSEUS MINING LIMITED — Capital/Financing Update 2017

Mar 30, 2017

31 March 2017

==> picture [195 x 60] intentionally omitted <==

NEWS RELEASE

PERSEUS UPDATES LIFE OF MINE PLAN FOR SISSINGUÉ GOLD MINE

Perseus Mining Limited (“Perseus” or the “Company”) (ASX/TSX: PRU) is pleased to announce details of its updated Life of Mine Plan (“LOMP”) for the Sissingué Gold Mine in Côte d’Ivoire, West Africa (“SGM”).

HIGHLIGHTS

• Perseus has updated its LOMP for the SGM assuming that re-estimated Mineral Resources and Ore Reserves for the Sissingué mineral deposit (“Sissingué”) as well as estimated Mineral Resources and Ore Reserves for the nearby Bélé East and Bélé West mineral deposits are processed through the SGM processing facility. The combined Proved and Probable Ore Reserves for three deposits total 5.9 million tonnes of ore, grading 2.1 g/t gold and containing 400 kozs ounces of gold;

• SGM’s estimated life of mine gold production totals 358 kozs including approximately 80,000 ounces/annum for the first 3.25 years of production and approximately 70,000 ounces/annum over the full 5 year life of mine;

• Forecast average weighted all-in site costs (“AISC”), including all direct production costs, royalties, waste stripping costs and sustaining capital expenditure, are estimated at approximately US$624 per ounce in the first 3.25 years of production and approximately US$628 per ounce over the full life of mine;

• The total capital cost estimate for the development of the SGM is US$107 million and the forecast sustaining capital costs (including the cost of site rehabilitation) which are included in the estimate of the AISC, total US$13 million;

• The SGM LOMP forecasts strong positive after tax cash flow totalling approximately US$104 million (or approximately A$0.13 per share at an A$:US$ exchange rate of 0.75), assuming a flat spot gold price of US$1,200 per ounce for unhedged ounces over the life of the mine starting from 1 March 2018 and assuming existing designated hedges for 67 kozs at a weighted average price of US$1,301 per ounce;

• Based on a total estimated construction cost of US$107 million, the ungeared, after tax internal rate of return from the project is approximately 28% (real) and the capital payback period is estimated at 39 months;

• Construction of the SGM is currently approximately 40% complete and is running on schedule and slightly under budget. Development is fully funded to completion through a combination of internal cash of US$67 million and bank provided project debt finance of US$40 million;

• First gold production at the SGM is expected to occur in the March Quarter of 2018.

• Recent exploration drilling on both the Sissingué Mining Lease and the nearby Mahalé Exploration Licence (all within trucking distance of the SGM processing facility) have demonstrated the potential for further increases in the inventory of material able to be processed at the SGM.

________________ Perseus Mining Limited ABN 27 106 808 986 Level 2, 437 Roberts Road Subiaco WA 6008 PO Box 1578 Subiaco WA 6008 Telephone: +61 8 6144 1700 Facsimile: +61 8 6144 1799 Email: info@perseusmining.com Website: www.perseusmining.com

==> picture [120 x 37] intentionally omitted <==

1. Overview

The updated LOMP for the SGM involves mining and processing of ore from three open pit mining areas based on the Sissingué, Bélé East and Bélé West mineral deposits as shown below in Figure 1.

==> picture [484 x 276] intentionally omitted <==

Figure 1: SGM layout

Pit optimisation, design and scheduling used a gold price of US$1,200 per ounce and input parameters based on Perseus’s operating experience including costs from recently contracted supply contracts. Based on the LOMP, the key forecast operating statistics for the SGM are summarised below in Table 1 .

Table 1: Overview Key Parameters of the updated LOMP

Parameter	Units	Average peryear	Total
Ex-Pit Mining		FY18-211,2	Life of Mine
Total ore + waste mined	Mt	6.7	25.0
Waste mined	Mt	5.1	19.2
Ore mined	Mt	1.6	5.8
Head grade	g/t gold	2.1	2.1
Strip ratio	t:t	3.2	3.3
Processing
Quantity ore processed	Mt	1.3 2.2 91 90.0 82	5.8 2.1 400 89.6 358
Head grade processed	g/t gold
Contained gold	‘000 ounces
Gold recovery rate	%
Gold production	‘000 ounces

2 | P a g e

==> picture [120 x 37] intentionally omitted <==

Table 1: Overview Key Parameters of the updated LOMP (Continued)

Parameter	Units	Average peryear	Total
Operating and Capital Costs		FY18-211,2	Life of Mine
Average mining costs	US$/tonne mined	3.29 12.27 6.33 541 50 33 624	3.17 13.16 6.71 543 50 35 628
Average processing costs	US$/tonne processed
Average general & administration (“G&A) costs	US$/tonne processed
Production costs	US$/ounce
Royalty	US$/ounce
Sustaining capital	US$/ounce
All-in site costs	US$/ounce

Notes:

1. Perseus has a financial year that ends on 30 June.

2. Covers the 3.25 year period from 1 April 2018 to 30 June 2021

2. Comparison of current updated LOMP relative to the previous LOMP

The current updated LOMP for the SGM differs from the previous version of the LOMP that was published in April 2015, as shown below in Table 2 .

Table 2: Overview Key Parameters

Parameter	Units	Current	Previous	% Change
		LOMP	LOMP
Mining
Total ore + waste mined	Mt	25.0	23.2	8
Waste mined	Mt	19.2	17.7	8
Ore mined	Mt	5.8	5.5	6
Head grade	g/t gold	2.1	2.4	(12)
Strip ratio	t:t	3.3	3.2	3
Processing
Quantity ore processed	Mt	5.8	5.5	6
Head grade processed	g/t gold	2.1	2.4	(12)
Contained gold	‘000 ounces	400	429	(7)
Gold recovery rate	%	89.6	89.7	0
Gold production	‘000 ounces	358	385	(7)
Average gold production	‘000 ounces/ year	71.6	73.4	(2)
Operating and Capital Costs
Average mining costs	US$/tonne mined	3.17	3.70	(14)
Average processing costs	US$/tonne processed	13.16	16.75	(21)
Average Site G&A costs	US$/tonne processed	6.71	7.70	(13)
Production costs	US$/ounce	543	569	(5)
Royalty	US$/ounce	50	49	(2)
Sustaining capital	US$/ounce	35	14	150
All-in site costs	US$/ounce	628	632	(1)

3. Mineral Resources and Ore Reserves

The updated global Measured and Indicated Mineral Resource for the SGM that combines the Sissingué and Bélé East and West deposits is now estimated as 15.0 Mt grading at 1.7 g/t gold, containing 820 kozs of gold

3 | P a g e

==> picture [120 x 37] intentionally omitted <==

as shown in Table 3. A further 1.4 Mt of material grading at 1.9 g/t gold and containing a further 80 kozs of gold are classified as Inferred Resources. Details of these estimates are shown below in Table 4 .

Table 3: Measured and Indicated Mineral Resources – March 2017

Deposit	Deposit Type	Measured Resources	Indicated Resources	Measured + Indicated Resources
		Quantity Grade Gold	Quantity Grade Gold	Quantity Grade Gold
		Mt g/t gold koz	Mt g/t gold koz	Mt g/t gold koz
Sissingué¹	Oxide Transitional Fresh	0.78 1.9 48 0.59 1.9 36 2.90 2.1 200	2.40 1.3 100 0.67 1.4 29 5.90 1.5 280	3.20 1.5 150 1.30 1.6 66 88.0 1.7 480
Sub-Total		4.30 2.1 290	8.90 1.4 410	13.00 1.6 700
Bélé East²	Laterite Completely weathered Partially weathered Weakly weathered Fresh	- - - - - - - - - - - - - - -	0.03 1.9 2 0.08 2.1 5 0.05 1.9 3 0.12 2.0 8 0.36 2.6 30	0.03 1.9 2 0.08 2.1 5 0.05 1.9 3 0.12 2.0 8 0.36 2.6 30
Sub-Total		- - -	0.65 2.3 49	0.65 2.3 49
Bélé West²	Laterite Completely weathered Partially weathered Weakly weathered Fresh	- - - - - - - - - - - - - - -	0.04 1.7 2 0.08 1.7 4 0.09 1.6 5 0.19 1.6 10 0.87 2.0 57	0.04 1.7 2 0.08 1.7 4 0.09 1.6 5 0.19 1.6 10 0.87 2.0 57
Sub-Total		- - -	1.300 1.8 78	1.30 1.8 78
Grand Total		4.30 2.1 290	11.00 1.5 540	15.00 1.7 820

Table 4: Inferred Mineral Resources – March 2017

Deposit	Deposit Type	Inferred Resources	Notes on Tables 3 and 4: 1. 0.6g/t gold cut-off applied at Sissingué 2. 0.8g/t gold cut-off grade applied at Bélé 3. Mineral Resources are inclusive of any Ore Reserves 4. Numbers are rounded and may not add up correctly
		Quantity Grade Gold
		Mt g/t gold koz
Sissingué	Oxide Transitional Fresh	0.23 1.2 9 0.06 1.0 2 0.66 2.3 48
Sub-total		0.94 1.9 58
Bélé East	Laterite Completely weathered Partially weathered Weakly weathered Fresh	0.02 1.5 1 0.01 1.9 1 - - - 0.02 1.5 1 0.24 1.8 14
Sub-total		0.28 1.8 16
Bélé West	Laterite Completely weathered Partially weathered Weakly weathered Fresh	0.02 2.0 1 0.02 1.7 1 0.01 1.3 1 0.02 1.6 1 0.08 2.0 5
Sub-total		0.14 1.8 8
Grand Total		1.400 1.9 80

4 | P a g e

==> picture [120 x 37] intentionally omitted <==

The SGM’s updated Ore Reserve which is summarised below in Table 5 is estimated at 5.9 million tonnes of ore, grading 2.1 g/t gold and containing 400 kozs of gold and is based on the re-estimated Sissingué Mineral Resource and the Bélé East and West Mineral Resource estimates as at February 2017 and updated pit optimisation, design and scheduling of the open pit resources. Table 5 reports the Ore Reserves by category, deposit and type, above variable cut-off grades. The classification categories of Proved and Probable under the JORC Code are equivalent to the CIM categories of the same name (CIM, 2010).

Table 5: Proved and Probable Ore Reserves - 31 March 2017

Deposit	Deposit Type	Proved	Probable	Proved + Probable
		Quantity Grade Gold	Quantity Grade Gold	Quantity Grade Gold
		Mt g/t gold koz	Mt g/t gold koz	Mt g/t gold koz
Sissingué3,5 Bélé East4,5 Bélé West3,4	Open pit Open pit Open pit	3.1 2.4 240 - - - - - -	1.8 1.5 86 0.5 2.5 39 0.5 2.1 35	4.8 2.1 330 0.5 2.5 39 0.5 2.1 35
Total		3.1 2.4 240	2.8 1.8 160	5.9 2.1 400

Notes:

1. Numbers are rounded and may not add up correctly in the table

2. All the estimates are on a dry tonne basis

3. Based on December 2016 Mineral Resource Estimate

4. Based on February 2017 Mineral Resource estimation

5. Variable gold cut-off grade based on material type and mining method

6. Inferred Mineral Resource is treated as mineralised waste

Proven and Probable Ore Reserves are found within the economic limits of three discrete open pit mining areas that have been designed based on Measured and Indicated Mineral Resources that incorporated all available Resource drilling results, a gold price of US$1,200 per ounce and mining, processing and general and administration costs derived from recent operating experience.

Economic assumptions used for Ore Reserve Estimation

• a. Gold metal price US$1,200 per ounce.

• b. A discount rate of 10% (real) has been assumed to calculate net present values of forecast cash flows unless specified otherwise.

• c. Un-escalated average costs used in optimising pit designs included:

Table 6: Operating costs (US$)

Deposit	Mining	Processing1,5	G&A	**Selling cost2 **	Royalty
Sissingué3,4 Bélé East3,4 Bélé West3,4	3.22/t mined 2.96/t mined 2.96/t mined	14.77/t processed 16.91/t processed 16.91/tprocessed	6.24/t processed 6.85/t processed 6.85/tprocessed	1.00 per ounce sold 1.00per ounce sold 1.00per ounce sold	4.5% 4% 4%

Notes:

1. Processing cost per tonne includes selling costs.

2. Selling costs include bullion transportation and refining.

3. Royalty includes amounts paid to the State (3.5%) and community development fund (0.5%) for all deposits, and Franco Nevada (0.5%) and Ivorian parties (US$0.80 per ounce) on the Sissingué pit only.

4. Government royalty operates on a sliding scale as follows:

5 | P a g e

==> picture [120 x 37] intentionally omitted <==

• a. 3.0%, where the spot sale price per ounce of gold is less than or equal to US$1,000;

• b. 3.5%, where the spot sale price per ounce of gold is more than US$1,000 and less than or equal to US$1,300;

• c. 4.0%, where the spot sale price per ounce of gold is more than US$1,000 and less than or equal to US$1,600;

• d. 5.0%, where the spot sale price per ounce of gold is more than US$1,000 and less than or equal to US$2,000;

• e. 6.0%, where the spot sale price of an ounce of gold is more than US$2,000s.

5. The assumed processing cost per tonne is lower than the processing cost per tonne in Table 2 due to recent tendering processes for key commodities subsequent to completion of the Ore Reserve estimates.

Mining parameters

• a. At Sissingué and Bélé, the chosen method of mining is conventional open pit mining utilising hydraulic excavators and trucks, mining bench heights of 5 m in ore and 10m in waste with 2.5 m flitches to minimise ore loss and waste rock dilution.

• b. The Sissingué optimisation cut-off grade calculations and pit designs were based on a gold price of US$1,150 per ounce, while the equivalent Bélé cut-off grades were based on US$1,200 per ounce. All financial evaluations were completed at US$1,200 per ounce.

• c. The economic pit shells were defined using Whittle 4X pit optimisation software (“Whittle 4X”) with inputs such as geotechnical parameters, ore loss and dilution, metallurgical recovery and mining costs.

• d. The pit optimisations were run with revenue generated only by Measured and Indicated Mineral Resources. No value was allocated to Inferred Mineral Resources.

• e. Whittle 4X input parameters were a combination of Perseus’s operating site experience at Edikan, supporting technical studies and contracts established specifically for Sissingué.

• f. Geotechnical parameters for Sissingué vary depending on the material type and Pit Sector. The inter-ramp slope angles are between 38 to 53 degrees. Similarly slope angles at Bélé vary between 38 and 56 degrees.

• g. Appropriate mining modifying factors such as ore loss, dilution and design parameters were used to convert the Mineral Resource to an Ore Reserve.

• h. At Sissingué, a smallest mining unit (“ SMU”) of 5.0 m east x 5.0 m west x 5.0 m high was selected resulting in approximately 9% ore loss and 15% mining dilution. At Bélé East the SMU of 2.5 m east x 5 m west x 5 m high was selected resulting in approximately 3% ore loss and 5% mining dilution. For Bélé West the SMU is 5 m east x 2.5 m west x 5 m high was selected, resulting in 6% ore loss and 10% mining dilution.

• i.

• Minimum mining width of 40 m was generally applied to the pit designs.

• j. As the mine is a green fields site, all surface infrastructure is required to enable the aforementioned mining method to be successfully implemented.

• k. There are no physical constraints to mining within the lease areas. No property, infrastructure or environmental issues are known to exist which may limit the extent of mining within the mining lease.

• l. Ore cut-off grades are based on the economic and mining parameters described above and are as follows:

Table 7: Cut-off Grades

Deposit		Cut-Off Grade by Ore Type (g/t gold)	Cut-Off Grade by Ore Type (g/t gold)	Cut-Off Grade by Ore Type (g/t gold)
	Oxide/Laterite	Transition	Granite/ Porphyry	Sediment	Mafic
Sissingué Bélé East Bélé West	0.45 0.65 0.65	0.85 1.00 1.00	0.85 1.05 1.05	1.05 NA NA	NA 1.20 1.20

6 | P a g e

==> picture [120 x 37] intentionally omitted <==

Processing Parameters

• a. The Sissingué processing plant will use crushing, grinding, gravity recovery and cyanide leaching to extract gold. The plant will have a nameplate throughput capacity of 1.8Mtpa on Sissingué oxide ore and 1.7Mtpa on Bélé oxide ore and 1.0Mtpa on fresh ore. Sissingué ore will be processed before Belé ore. Sissingué ore will have a grind size of 106μm and Bélé has a grind size of 75μm.

• b. The technology to be used in the processing plant is well proven in many other operations globally.

• c. The processing test work is representative of the different material types throughout the mining areas.

• d. No deleterious material has been identified.

• e. Metallurgical test work has been carried out to a standard that is considered representative of the orebodies as a whole.

• f. The process metallurgical recovery for gold is fixed by material type.

Table 8: Recoveries

Deposit			Cut-Off Grade by Ore Type (g/t gold)	Cut-Off Grade by Ore Type (g/t gold)	Cut-Off Grade by Ore Type (g/t gold)
	Oxide/Laterite		Transition	Granite/ Porphyry	Sediment	Mafic
Sissingué Bélé East Bélé West	91 94 94		91 93 93	90 91 91	83¹ NA NA	NA 91 91

Notes:

1. Average recovery based on the formula of log (Au x 7.63) + 78.5.

Classification

• a. Ore Reserves have been classified based on the underlying Mineral Resource classifications and the level of detail in the mine planning. The Mineral Resources were classified as Measured, Indicated and Inferred. The Ore Reserves, based only on the Measured and Indicated Resources, have been classified as Proved and Probable Ore Reserves, respectively.

• b. The Ore Reserve is classified as Proved and Probable in accordance with the JORC Code, corresponding to the Mineral Resource classifications of Measured and Indicated and taking into account other factors where relevant. Each deposit’s geological model is well constrained. The Ore Reserve classification is considered appropriate given the nature of the deposits, the moderate grade variability, drilling density, structural complexity and mining history. Therefore it was deemed appropriate to use Measured Mineral Resources as a basis for Proved Reserves and Indicated Mineral Resources as a basis for Probable Reserves.

• c.

• No Inferred Mineral Resources were included in the Ore Reserve estimate.

7 | P a g e

==> picture [120 x 37] intentionally omitted <==

4. Gold production profile

With the aim of maximising the return on funds employed at the SGM, the mining sequence of the pits along with mill feed profile has been optimised and scheduled. The result of this scheduling is that the gold production profile and resulting cash flows from the SGM remain strong for the remainder of the mine life. (Refer to Figures 2, 3 and 4 below.)

Figure 2: Sissingué Gold Mine - scheduled monthly material movement ex-pit (tonnes)

==> picture [474 x 144] intentionally omitted <==

Figure 3: Sissingué Gold Mine – Monthly tonnes and grade to Mill

==> picture [474 x 144] intentionally omitted <==

Figure 4: Sissingué Gold Mine – Gold Production by Pit

==> picture [474 x 144] intentionally omitted <==

8 | P a g e

==> picture [120 x 37] intentionally omitted <==

5. Future cash flows and sensitivity analysis

Based on the LOM gold production and cost parameters, the net after tax cash flows forecast to be generated by the SGM from 1 March 2018, at US$1,200 per ounce gold price, are estimated to total US$104 million or approximately A$0.13 per share (assuming an A$:US$ exchange rate of 0.75).

The sensitivity analysis shown below in Table 9 summarises the sensitivity of the SGM’s net after tax cash flows to movements in the gold price. Within an expected short term trading range of US$1,100 per ounce to US$1,300 per ounce, the projected cash flows of the mine remain robust and together with strong cash flows from Perseus’s producing Edikan Gold Mine in Ghana, is capable of materially contributing to the development funding of Perseus’s third project, namely the Yaouré Gold Mine, also located in Côte d’Ivoire.

Table 9: Sensitivity of SGM’s After Tax Cash flow and IRR to gold price movements

		Gold Price	Gold Price	Gold Price
		US$1,100/ounce	US$1,200/ounce	US$1,300/ounce
After Tax Cash Flow Internal Rate of Return (IRR) Payback Period	USD million % Months	77 21 40	104 28 39	130 33 38

To discuss any aspect of this announcement, please contact:

Managing Director: Jeff Quartermaine at telephone +61 8 6144 1700 or email jeff.quartermaine@perseusmining.com; Investor Relations: Cathy Moises at telephone + 61 412196350 or email cathy.moises@perseusmining.com (Perth/Melbourne);

Media Relations: Nathan Ryan at telephone +61 4 20 582 887 or email nathan.ryan@nwrcommunications.com.au (Melbourne)

Caution Regarding Forward Looking Information :

This report contains forward-looking information which is based on the assumptions, estimates, analysis and opinions of management made in light of its experience and its perception of trends, current conditions and expected developments, as well as other factors that management of the Company believes to be relevant and reasonable in the circumstances at the date that such statements are made, but which may prove to be incorrect. Assumptions have been made by the Company regarding, among other things: the price of gold, continuing commercial production at the Edikan Gold Mine without any major disruption, development of a mine at Tengrela, the receipt of required governmental approvals, the accuracy of capital and operating cost estimates, the ability of the Company to operate in a safe, efficient and effective manner and the ability of the Company to obtain financing as and when required and on reasonable terms. Readers are cautioned that the foregoing list is not exhaustive of all factors and assumptions which may have been used by the Company. Although management believes that the assumptions made by the Company and the expectations represented by such information are reasonable, there can be no assurance that the forward-looking information will prove to be accurate. Forward-looking information involves known and unknown risks, uncertainties, and other factors which may cause the actual results, performance or achievements of the Company to be materially different from any anticipated future results, performance or achievements expressed or implied by such forward-looking information. Such factors include, among others, the actual market price of gold, the actual results of current exploration, the actual results of future exploration, changes in project parameters as plans continue to be evaluated, as well as those factors disclosed in the Company's publicly filed documents. The Company believes that the assumptions and expectations reflected in the forward-looking information are reasonable. Assumptions have been made regarding, among other things, the Company’s ability to carry on its exploration and development activities, the timely receipt of required approvals, the price of gold, the ability of the Company to operate in a safe, efficient and effective manner and the ability of the Company to obtain financing as and when required and on reasonable terms. Readers should not place undue reliance on forward-looking information. Perseus does not undertake to update any forward-looking information, except in accordance with applicable securities laws.

9 | P a g e

==> picture [120 x 37] intentionally omitted <==

Competent Person Statement

All production targets for the Sissingué Gold Mine referred to in this report are underpinned by estimated Ore Reserves which have been prepared by competent persons in accordance with the requirements of the JORC Code. The information in this report that relates to Mineral Resources for Sissingué was first reported by the Company in compliance with the JORC Code 2012 and NI43-101 in a market announcement released on 15 December 2016. The information in this report that relates to Mineral Resources for Bélé was first reported by the Company in compliance with the JORC Code 2012 and NI43-101 in a market announcement released on 20 February 2017. The Company confirms that it is not aware of any new information or data that materially affects the information in those market announcements and that all material assumptions and technical parameters underpinning the estimates in that market announcement continue to apply and have not materially changed. The Company further confirms that material assumptions underpinning the estimates of Ore Reserves described in “Technical Report — Sissingué Gold Project, Côte d’Ivoire” dated 29 May 2011 continue to apply.

Mr Joe McDiarmid , who is a Chartered Professional Member of the Australasian Institute of Mining and Metallurgy, and is an employee of RungePincockMinarco Limited has compiled and reviewed the information in this release and Appendix 1 which relate to the Sissingué Ore Reserves. Mr Joe McDiarmid has sufficient experience, which is relevant to the style of mineralisation and type of deposit under consideration and to the activity which he has undertaken to qualify as a Competent Person, as defined in the JORC Code 2012 and a Qualified Person as defined NI 43-101. Mr McDiarmid has no economic, financial or pecuniary interest in the Company and consents to the inclusion in this report of the matters based on this information in the form and context in which it appears and has approved the inclusion of technical and scientific information in this report.

Mr Paul Thompson , who is a Fellow of the Australasian Institute of Mining and Metallurgy and is an employee of Perseus Mining Limited has compiled and reviewed the information in this release and Appendix 2 which relates to the Bélé Ore Reserves. Mr Thompson has sufficient experience, which is relevant to the style of mineralisation and type of deposit under consideration and to the activity which he has undertaken to qualify as a Competent Person as defined in the JORC Code 2012 and a Qualified Person as defined in NI43-101 and consents to the inclusion in this report of the matters based on this information in the form and context in which it appears and has approved the inclusion of technical and scientific information in this report.

10 | P a g e

==> picture [120 x 37] intentionally omitted <==

1. Appendix 1: JORC Table 1 For Sissingué

Section 1 to 3 of Table 1 has been prepared by Snowden and can be referenced as part of the 2017 Statement of Mineral Resources for Sissingué.

1.1 Sissingué JORC Code 2012 Section 1 Sampling Techniques and Data

This section has been prepared by Snowden to support the Statement Mineral Resources for Sissingué as of 15[th] December 2016.

Criteria	JORC Code explanation	Commentary
Sampling techniques	 Nature and quality of sampling (e.g. cut channels, random chips, or specific specialised industry standard measurement tools appropriate to the minerals under investigation, such as downhole gamma sondes, or handheld XRF instruments, etc.). These examples should not be taken as limiting the broad meaning of sampling.  Include reference to measures taken to ensure sample representivity and the appropriate calibration of any measurement tools or systems used.  Aspects of the determination of mineralisation that are Material to the Public Report.  In cases where ‘industry standard’ work has been done this would be relatively simple (e.g. ‘reverse circulation drilling was used to obtain 1 m samples from which 3 kg was pulverised to produce a 30 g charge for fire assay’). In other cases more explanation may be required, such as where there is coarse gold that has inherent sampling problems. Unusual commodities or mineralisation types (e.g. submarine nodules) may warrant disclosure of detailed information.	 Sissingué data used for estimating the Mineral Resource includes: o 1,654 RC drill holes for 122,889 m. o 379 diamond drill holes for 77,055 m. o 18 diamond drill holes with RC pre-collars for 2,163 m.  RC drilling (5¼” diameter) was used to collect 1 m samples. Majority composited to 2 m samples (by weighing); close spaced infill submitted as 1 m samples. Sample weight nominally of 2.5 kg and 5 kg respectively.  Half-core from diamond drill holes (‘right’ side systematically taken; 1.5 m in oxide and transition, 1 m in fresh).  50 g charge produced for fire assay.
Drilling techniques	 Drill type (e.g. core, reverse circulation, open-hole hammer, rotary air blast, auger, Bangka, sonic, etc.) and details (e.g. core diameter, triple or standard tube, depth of diamond tails, face- sampling bit or other type, whether core is oriented and if so, by what method, etc.).	 RC drilling (5¼” diameter), usually 80 m or less in depth. Generally RC holes have collar azimuth and inclination only measured.  Diamond drilling, HQ in weathered rock, NQ in fresh rock. All diamond holes downhole surveyed at 30 m intervals.  43 holes oriented by core spear; 217 holes oriented by“AceTool”device.
Drill sample recovery	 Method of recording and assessing core and chip sample recoveries and results assessed.  Measures taken to maximise sample recovery and ensure representative nature of the samples.  Whether a relationship exists between sample recovery and grade and	 1 m RC samples weighed and composited to 2 m.  Length of recovered diamond core measured and recovery calculated based on run length; close to 100% recovery for virtually all core in fresh rock.  There is no apparent relationship between sample recovery and grade for diamond drilling.

11 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
	whether sample bias may have occurred due to preferential loss/gain of fine/coarse material.
Logging	 Whether core and chip samples have been geologically and geotechnically logged to a level of detail to support appropriate Mineral Resource estimation, mining studies and metallurgical studies.  Whether logging is qualitative or quantitative in nature. Core (or costean, channel, etc.) photography.  The total length and percentage of the relevant intersections logged.	 RC drill chip boards were prepared and the chips logged geologically, including rock type, alteration type and intensity (where recognisable), vein quartz content in estimated percentage, sulphide mineralisation and estimated content and weathering domain.  Diamond drill core was geologically and structurally logged and photographed, before being sawn in half, including fault, fold, cleavage and joint orientation, lithological contacts, vein orientation and bedding. Logged items are lithology, weathering, colour, grain size, vein type and vein volume percentage, sulphide mineralisation and their estimated percentage, alteration and alteration intensity.  All sample intervals in both RC and diamond holes were sampled and assayed.  Logging is considered appropriate and reliable.
Sub-sampling techniques and sample preparation	 If core, whether cut or sawn and whether quarter, half or all core taken.  If non-core, whether riffled, tube sampled, rotary split, etc. and whether sampled wet or dry.  For all sample types, the nature, quality and appropriateness of the sample preparation technique.  Quality control procedures adopted for all sub-sampling stages to maximise representivity of samples.  Measures taken to ensure that the sampling is representative of the in situ material collected, including for instance results for field duplicate/second-half sampling.  Whether sample sizes are appropriate to the grain size of the material being sampled.	 RC samples were collected at the drill site at 1 m interval and split using a multi-stage riffle splitter. Each two consecutive samples were composited (where applicable) in one bag. Wet samples were sub-sampled using a spear.  Diamond core sawn in half using a motorized diamond blade saw; right half sent for assaying, left half stored in core trays for reference.  Both core and RC samples followed a standard path of drying, crushing and grinding. Samples were pulverised with a ring mill and thoroughly mixed on a rolling mat (“carpet roll”), and then 200 g of sub- sample was collected. Internal laboratory checks required at least 90% of the pulp passing -75 µm.  Some RC samples at depth were identified as having downhole contamination and resultant smearing of grades as a result of wet drilling in ‘sticky’ material, with the samples being ‘hung up’ in the cyclone and subsequently contaminating later samples. As a result of this, all RC holes in the pit area were reviewed and any suspected of containing smeared assays were removed from the dataset prior to estimation. Approximately 5% of RC samples were removed due to suspected downhole contamination. In addition, 2016 drilling focused on diamond drill holes to confirm areas with RC drilling in the core of the deposit.  With the exception of the issue noted above, the sub-sampling is considered appropriate and representative.
Quality of assay data and laboratory tests	 The nature, quality and appropriateness of the assaying and laboratory procedures used and whether the technique is considered partial or total.  For geophysical tools, spectrometers, handheld XRF instruments, etc., the parameters used in determining the analysis including instrument make	 A small number of initial holes were assayed by bottle roll, which was found to be partial and inaccurate.  All subsequent assaying by standard Fire Assay.  Field duplicates (RC only) inserted at 1 in 25.  No field duplicates for DD as ¼ core considered as inadequate sample, and submission remaining ½ core considered undesirable.

12 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
	and model, reading times, calibrations factors applied and their derivation, etc.  Nature of quality control procedures adopted (e.g. standards, blanks, duplicates, external laboratory checks) and whether acceptable levels of accuracy (i.e. lack of bias) and precision have been established.	 Blanks inserted at 1 in 25.  Certified standards at 1 in 50 up to 2008; thereafter at 1 in 20.  Internal laboratory standards, duplicates and repeats and various other tests have been carried out throughout the drilling programs.  QAQC shows no bias, but only moderate reproducibility, particularly at high grades. This is as expected with the nugget mineralisation.  Overall assaying quality is considered acceptable with the exception of the potential smearing in some RC samples which were subsequently removed from the dataset prior to estimation as previously discussed.
Verification of sampling and assaying	 The verification of significant intersections by either independent or alternative company personnel.  The use of twinned holes.  Documentation of primary data, data entry procedures, data verification, data storage (physical and electronic) protocols.  Discuss any adjustment to assay data.	 During the 2016 drilling program, Perseus drilled a number of diamond drill holes to confirm the grade tenor and check RC drill holes suspected of downhole contamination and smearing. As a result of this program, approximately 5% of RC samples were removed from the dataset where the RC grades were not supported by the diamond.  Drill hole information for both RC and diamond core holes is captured at the drill site on paper.  All hard copies are handed over to the database administrator in Tengréla site office and the information provided on paper is entered into the computer.  All hard copies are kept in Tengréla site office.  Downhole survey data and collar survey data are provided by the drilling contractors and surveyors respectively in digital format.  Perseus maintains a centralised database for its various operations in Ghana and Côte d’Ivoire. Database administration is based in Perseus’ head office in Accra/Ghana and under the supervision of the company’s Resource Geologist.  No adjustments were made to the raw assay data with the exception of the removal of any RC samples with suspected smearing of grades as previously discussed. Top cutting is only applied after database compositing and statistical analysis and prior to resource estimation.
Location of data points	 Accuracy and quality of surveys used to locate drill holes (collar and downhole surveys), trenches, mine workings and other locations used in Mineral Resource estimation.  Specification of the grid system used.  Quality and adequacy of topographic control.	 All RC and diamond holes were surveyed using differential GPS, until September 2009 by a certified contract surveyor (SEMS Exploration Services Ltd, Ghana). Drill holes between September 2009 and October 2010 were surveyed by CBM Surveys Ltd of Ghana. All subsequent drill holes were surveyed by the company’s surveyor.  Grid system used is WGS84 UTM 29N.  The topography covering the extent of the Sissingué Mineral Resource model was created as a digital terrain model (DTM) in Surpac using the accurately- surveyed drill hole collar data and an additional 639 survey points across the prospect.
Data spacing and	 Data spacing for reporting of Exploration Results.  Whetherthe data spacing and	 Data spacing for resource estimation varies from 10 m x 10 m to 20 m x 20 m for most areas of the deposit.

13 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
distribution	distribution is sufficient to establish the degree of geological and grade continuity appropriate for the Mineral Resource and Ore Reserve estimation procedure(s) and classifications applied.  Whether sample compositing has been applied.	 Where data spacing is wider (to a maximum of 40 m x 40 m), an Inferred classification is used.  Data spacing is sufficient to establish grade and geological continuity appropriate to the resource estimation procedures and classifications applied.  Samples have been composited (by computer) to 2 m, honouring geological divisions.
Orientation of data in relation to geological structure	 Whether the orientation of sampling achieves unbiased sampling of possible structures and the extent to which this is known, considering the deposit type.  If the relationship between the drilling orientation and the orientation of key mineralised structures is considered to have introduced a sampling bias, this should be assessed and reported if material.	 Orientation of drilled section lines is dominantly at right angles to the strike of the geology and mineralisation domains.  Drillholes are angled to cross the sub-vertical dip of the geological domains.  Disseminated mineralised veins have developed within the overall geological domains; the estimation method is considered to account for this.
Sample security	 The measures taken to ensure sample security.	 Samples from RC drilling are collected and bagged at drill site during the drilling operation. Core samples are cut in a central facility in Tengréla and samples placed into sample bags as they are cut.  All samples are then catalogued and placed in large woven bags and sealed prior to dispatch to ALS, Intertek or BVML for preparation and analysis.  Dispatch from site to Korhogo (Intertek) is undertaken by Perseus staff and vehicles.  Samples dispatched to ALS and BVML are collected from Tengréla by staff and vehicles of the respective laboratories.  All aspects of the process are supervised by Perseus personnel and limited opportunity exists for tampering with samples.
Audits or reviews	 The results of any audits or reviews of sampling techniques and data.	 Steffen Brammer of Perseus has reviewed sample techniques and data during regular site visits between 2008 and 2013, and considers them adequate.  Reviews were also carried out by personnel from consulting company Runge Limited during 2009 and 2010, and Mr L Widenbar of Widenbar & Associates in October 2012 with acceptable conclusions.  Snowden visited site during December 2016 and reviewed the drilling and sampling procedures being carried out at the nearby Bélé deposit. These procedures are the same as those used for the latest Sissingué drilling program and Snowden considers them adequate. For the 2016 drilling, where RC drilling is used, the RC drilling is conducted until wet samples are returned, at which point the drilling is changed to diamond drilling to avoid the contamination of RC samples.

14 | P a g e

==> picture [120 x 37] intentionally omitted <==

1.2 Sissingué JORC Code 2012 Section 2 Reporting of Exploration Results

This section has been prepared by Snowden to support the Statement Mineral Resources for Sissingué as of 15th December 2016.

Criteria	JORC Code explanation	Commentary
Mineral tenement and land tenure status	 Type, reference name/number, location and ownership including agreements or material issues with third parties such as joint ventures, partnerships, overriding royalties, native title interests, historical sites, wilderness or national park and environmental settings. The security of the tenure held at the time of reporting along with any known impediments to obtaining a licence to operate in the area.	 The Sissingué Mineral Resource lies within mining permit PE39 (Permit d’Exploitation Sissingué).  Perseus holds an 86% interest in PE39 through the Company’s wholly owned subsidiary Perseus Mining Côte d’Ivoire SA. The government of Côte d’Ivoire holds a 10% free carried interest in the property and the remaining 4% interest is held by local joint venture partner Société Minière de Côte d’Ivoire (SOMICI).  The mining permit is valid for 6 years until August 2018 and is renewable.  The Government of Côte d’Ivoire is entitled to a royalty on production as follows: . Spot price per ounce - London PM Fix Royalty Rate Less than or equal to US$1000 3% Higher than US$1000 and less than or equal to US$1300 3.5% Higher than US$1300 and less than or equal to US$1600 4% Higher than US$1600 and less than or equal to US$2000 5% Higher than US$2000 6%  Franco Nevada are entitled to a 0.5% royalty on production.  The Sissingué Project area has no known environmental liabilities.
Exploration done by other parties	 Acknowledgment and appraisal of exploration by other parties.	 All exploration was by Perseus using soil geochemical sampling, with follow-up drilling in areas with anomalous gold mineralisation, which led to the discovery of the Sissingué deposit.
Geology	 Deposit type, geological setting and style of mineralisation.	 The Sissingué Deposit occurs in a strongly deformed Birimian greenstone belt intruded by granitoid bodies.  Gold mineralisation at Sissingué is associated with the porphyritic dykes of tonalitic chemistry that cross cut the flysch sediments (turbidites).  Subsequenthydrothermalactivities and

15 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
		metasomatism of the tonalite has led to a sericite-carbonate alteration within the intrusives and the more permeable horizons (sandstones and conglomerates) of the turbidites, and a low to moderate grade disseminated gold mineralisation.  Late stage high grade Au-As-quartz- carbonate veins exploited the altered and brittle portions of the intrusives and sediments with common occurrences of visible gold.
Drill hole Information	 A summary of all information material to the understanding of the exploration results including a tabulation of the following information for all Material drill holes:  Easting and northing of the drill hole collar.  Elevation or RL (Reduced Level – elevation above sea level in metres) of the drill hole collar.  Dip and azimuth of the hole.  Downhole length and interception depth.  Hole length. If the exclusion of this information is justified on the basis that the information is not Material and this exclusion does not detract from the understanding of the report, the Competent Person should clearly explain why this is the case.	 Exploration results are not being reported for the Mineral Resource area.
Data aggregation methods	 In reporting Exploration Results, weighting averaging techniques, maximum and/or minimum grade truncations (e.g. cutting of high grades) and cut-off grades are usually Material and should be stated.  Where aggregate intercepts incorporate short lengths of high grade results and longer lengths of low grade results, the procedure used for such aggregation should be stated and some typical examples of such aggregations should be shown in detail.  The assumptions used for any reporting of metal equivalent values should be clearly stated.	 Exploration results are not being reported for the Mineral Resource area.
Relationship between mineralisation widths and intercept lengths	 These relationships are particularly important in the reporting of Exploration Results.  If the geometry of the mineralisation with respect to the drill hole angle is known, its nature should be reported. If it is not known and only the downhole lengths are reported, there should be a clear statement to this effect (e.g. ‘downhole length, true width not known’).	 Exploration results are not being reported for the Mineral Resource area.  Orientation of drilled section lines is dominantly at right angles to the strike of the geology and mineralisation domains.  Drillholes are angled to cross the sub- vertical dip of the geological domains.
Diagrams	Appropriate maps and sections (with scales) and tabulations of intercepts should be included for any significant discovery being reported These should include, but not be limited to a plan view of drill hole collar locations and	 Location plans and example sections are included in the Mineral Resource technical documentation.

16 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
	appropriate sectional views.
Balanced reporting	Where comprehensive reporting of all Exploration Results is not practicable, representative reporting of both low and high grades and/or widths should be practiced to avoid misleading reporting of Exploration Results.	 Exploration results are not being reported for the Mineral Resource area.
Other substantive exploration data	Other exploration data, if meaningful and material, should be reported including (but not limited to): geological observations; geophysical survey results; geochemical survey results; bulk samples – size and method of treatment; metallurgical test results; bulk density, groundwater, geotechnical and rock characteristics; potential deleterious or contaminating substances.	 Exploration results are not being reported for the Mineral Resource area.
Further work	 The nature and scale of planned further work (eg tests for lateral extensions or depth extensions or large-scale step-out drilling).  Diagrams clearly highlighting the areas of possible extensions, including the main geological interpretations and future drilling areas, provided this information is not commercially sensitive.	 Exploration results are not being reported for the Mineral Resource area.  A Feasibility Study has been carried out at Sissingué. The Ore Reserve and pit designs will be updated with this latest Mineral Resource. Exploration over possible satellite deposits is currently on- going.

1.3 Sissingué JORC Code 2012 Section 3 Estimation and Reporting of Mineral Resources

This section has been prepared by Snowden to support the Statement Mineral Resources for Sissingué as of 15th December.

Criteria	JORC Code explanation	Commentary
Database integrity	 Measures taken to ensure that data has not been corrupted by, for example, transcription or keying errors, between its initial collection and its use for Mineral Resource estimation purposes.  Data validation procedures used.	 Perseus maintains a centralised database for its various operations in Ghana and Côte d’Ivoire. Database administration is based in Perseus’ head office in Accra/Ghana and under the supervision of the company’s Resource Geologist.  Perseus carried out detailed validation of the dataset and retain overall responsibility for the database quality. All drill hole data was validated during data entry by Perseus including: - Checks for duplicate collars (LogChief, Datashed). - Checks for missing samples (Datashed). - Checks for down hole from-to interval consistency (LogChief, Datashed). - Checks for overlapping samples (LogChief, Datashed). - Checks for samples beyond hole depth (LogChief, Datashed). - Checks for inexistent or misspelt log items (LogChief). - Check for missing assays (Datashed). - Check fordown-holeinformationbeyondhole

17 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
		depth (Datashed).  Snowden carried out a basic statistical and visual validation prior to estimation including: - Drillholes with overlapping sample intervals. - Sample intervals with no assay data. - Duplicate records. - Assay grade ranges. - Collar coordinate ranges. - Valid hole orientation data.  No validation issues were found with the data and Snowden considers the data to be appropriate for estimation.  In addition, Snowden reviewed all RC samples within the pit area to assess them for potential downhole contamination and resultant smearing. Any samples suspected of containing smeared assays were removed from the dataset prior to estimation. This impacts approximately 5% of RC samples.
Site visits	 Comment on any site visits undertaken by the Competent Person and the outcome of those visits.  If no site visits have been undertaken indicate why this is the case.	 Steffen Brammer of Perseus has undertaken regular site visits between 2008 and 2013. Steffen has reviewed the geology and data collection processes during this time,  Lynn Olssen (Senior Principal Consultant) of Snowden Mining Industry Consultants visited site during December 2016. Lynn reviewed the drilling and sampling procedures for diamond and RC drilling at the nearby Bélé deposits. Lynn also visited the Sissingué site and core shed and reviewed the diamond core sampling processes and diamond core, RC chip boards, logging procedures and general site layout.
Geological interpretation	 Confidence in (or conversely, the uncertainty of ) the geological interpretation of the mineral deposit.  Nature of the data used and of any assumptions made.  The effect, if any, of alternative interpretations on Mineral Resource estimation.  The use of geology in guiding and controlling Mineral Resource estimation.  The factors affecting continuity both of grade and geology.	 Mineralisation was domained by its host lithology for modelling. The three main host lithologies are granite, porphyritic dykes and sediments. Mineralised sediments in the alteration halo of the dykes were included in the dyke domain to maintain a minimum width of the wireframes and to maintain continuity along strike.  Where geological contacts were not clearly controlling the distribution of mineralisation, a grade cut-off of 0.3 g/t Au was used to construct Mineral Resource boundaries and to provide overall geometry to mineralised zones. A minimum of 4 m width was used for the wireframes and samples of grades below the nominal cut-off of 0.3 g/t Au were included where the wireframe would otherwise be less than 4 m wide. Analysis of the global grade distribution shows that there is a natural change in grade population at around 0.3 g/t Au.  Geological continuity of the granite and sediment domains is understood with reasonable confidence. The classification reflects this level of confidence.  Porphyry lithologies (dyke domains) were limited to Inferred and Indicated categories due to the mostly narrow nature of the mineralised bodies and consequently the decreased confidence in their geometry.  Continuity and variability is also influenced by weathering and these have been interpreted and incorporated into the oxide, transitional and fresh domains.
Dimensions	 The extent and variability of the Mineral Resource expressed as length (along strike or otherwise), plan width, and depth below surface to the	 The Sissingué deposit extends for 3.15 km along strike and to a depth of typically 140 m to 160 m below surface, with a maximum depth of 290 m.  Thickness across strike is typically 50 m to 80 m for the granite and sediment domains, but limited to 10 m to 20 m

18 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
	upper and lower limits of the Mineral Resource.	for the porphyry dykes.
Estimation and modelling techniques	 The nature and appropriateness of the estimation technique(s) applied and key assumptions, including treatment of extreme grade values, domaining, interpolation parameters and maximum distance of extrapolation from data points. If a computer assisted estimation method was chosen include a description of computer software and parameters used.  The availability of check estimates, previous estimates and/or mine production records and whether the Mineral Resource estimate takes appropriate account of such data.  The assumptions made regarding recovery of by- products.  Estimation of deleterious elements or other non-grade variables of economic significance (e.g. sulphur for acid mine drainage characterisation).  In the case of block model interpolation, the block size in relation to the average sample spacing and the search employed.  Any assumptions behind modelling of selective mining units.  Any assumptions about correlation between variables.  Description of how the geological interpretation was used to control the resource estimates.  Discussion of basis for using or not using grade cutting or capping.  The process of validation, the checking process used, the comparison of model data to drill hole data, and use of reconciliation data if available.	Boundary conditions:  The granite, sediments and dykes were kept separate for statistical analysis, variography and estimation as they are different geological units with mostly hard boundaries between them. The southern and northern granite and sediment domains were also kept separate for estimation as there is a physical separation between the two areas. While there is a grade difference between the northern and southern areas, particularly for the granite (higher in the southern area), the grade distributions are of a similar shape and hence the two areas were combined for variography to provide a larger dataset for analysis.  Based on statistical analysis and boundary conditions, the oxide and transitional domains were combined for variography and estimation while the fresh domain was kept separate. Visual inspection of the drill hole grades between the laterite and the oxide does not show evidence of a depleted zone within the mineralised domains. As a result of the small dataset, with only minor areas of mineralisation in the lower parts of the laterite, the laterite was combined with the oxide for estimation. The laterite data was not used for variography.  Review of the individual dykes (and weathering domains) showed that they are statistically similar and, given the relatively small amount of data in the individual dykes, the dykes were combined for statistical analysis and variography. The dykes were kept separate for estimation as there are hard boundaries between them. Estimation – granites and sediments:  Due to the highly skewed nature and presence of mixed populations in the granites and sediments domains, multiple indicator kriging (MIK) was used to estimate gold grades. CAE Studio (Datamine) software was used to estimate the probability of the grade being above or below a series of thresholds into parent blocks of 10 mE by 10 mN by 5 mRL. Thresholds were defined for each estimation domain.  Kriging neighbourhood analysis (KNA) was used to define an appropriate block size for estimation and number of informing samples. Blocks were estimated using a minimum and maximum number of samples of 8 and 14 respectively. A maximum of 6 composites was allowed per drill hole for estimation.  The initial search pass used ranges equivalent to the ranges of continuity seen in the variograms at around 90% to 100% of the variance in the 0.3 g/ Au to 0.5 g/t Au threshold variograms, with the search ellipse orientated as per the higher grade thresholds (0.5 g/t Au and above).  Post processing of the MIK probability estimates was carried out in GSLIB software. Post processing was used to carry out order relation corrections using an averaging approach, and to calculate the e-type grade for each block. The e-type grades were calculated using a linear model between indicator thresholds and a hyperbolic or power model for the upper and lower tails respectively. Estimation – dykes:  For the lower grade dykes, ordinary kriging was selected as appropriate,withtop cuts tolimit theinfluence ofthehigher

19 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
		grade composites. A top cut of 30 g/t Au was selected for the dykes in all weathering domains, based on review of the histograms and natural breaks in the populations in the log- probability plots.  Ordinary kriging (parent cell estimates) was used to estimate gold grades for the dykes using a discretisation of 5 east by 5 north by 3 elevation.  Due to the positively skewed nature of the grade distributions, normal scores variograms were modelled for the dykes.  A dynamic anisotropy approach was used, whereby the true dip and azimuth of the mineralised lodes was estimated into each block in the model and the search and variogram orientations were locally adjusted to reflect the geological orientation. This method allows the estimate to better reflect the changing orientation and undulating nature of some of these dykes along strike.  Blocks were estimated using a minimum and maximum number of samples of 8 and 18 respectively based on the KNA work. A maximum of 6 composites was allowed per drill hole for estimation.  The initial search pass used ranges equivalent to the ranges of continuity seen in the variograms at around 90 % to 100% of the variance. Validation:  Snowden validated the Sissingué model by: - Global comparison of top cut and declustered input grades with tonnage weighted output grades for each domain. - Visual inspection of the model against the input composites. - Comparison of moving window input and output statistics. - Global change of support to check level of selectivity in model. - Snowden considers the estimate to be a reasonable reflection of the input data. o There has been no mining to date; no reconciliation data is available.
Moisture	 Whether the tonnages are estimated on a dry basis or with natural moisture, and the method of determination of the moisture content.	 Tonnages are reported on a dry basis.
Cut-off parameters	 The basis of the adopted cut-off grade(s) or quality parameters applied.	 The reporting cut-off is based on optimisation studies carried out as part of the 2010 Feasibility Study, which have suggested that the deposit can be economically extracted at a gold cut-off in the range 0.4 to 0.6 g/t.
Mining factors or assumptions	 Assumptions made regarding possible mining methods, minimum mining dimensions and internal (or, if applicable, external) mining dilution. It is always necessary as part of the process of determining reasonable prospects for eventual economic extraction to consider potential mining	 Trial optimisation was run on the 2014 Mineral Resource in Whittle at a USD2,400 gold price (approximately double the current spot price) to define the base of potentially economic open-pittable material. The same shell was used to constrain the 2016 Mineral Resource.  Assumptions for the optimisation were based on studies carried out by Perseus as follows.  Cut-off grades: o Oxide

20 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
	methods, but the assumptions made regarding mining methods and parameters when estimating Mineral Resources may not always be rigorous. Where this is the case, this should be reported with an explanation of the basis of the mining assumptions made.	0.8 g/t Au o Transition 1.0 g/t Au o Granite/Porphyry 1.1 g/t Au o Sediments 1.3 g/t Au  Whittle processing cost (treatment + general administration) were: o Oxide ($21.79/t+$7.19/t) $28.98/t o Transition ($25.04/t+$8.14/t) $33.18/t o Granite/Porphyry ($26.29/t+$8.32/t) $34.61/t o Sediments ($30.29/t+$9.87/t) $40.13/t  Dilution: o Mining ore recovery (before diluted) 97% o Mining ore dilution (at 0 g/t Au dilutant grade) 3%  Geotechnical parameters: o Oxide (approximately from 390 mRL to 325 mRL) 31 degrees o Transition (approximately from 325 mRL to 300mRL) 41 degrees o Fresh (approximately from 300 mRL and below) 45 degrees  Mining cost estimate: o Contract miner $3.61/t o Fuel ($1.40/L) $0.59/t o Day-works (3%) $0.13/t o Total mining cost $4.33/t  Whittle mining cost calculations: o Base cost at surface $4.10/t o Incremental trucking cost $0.005/t/vm  The Whittle optimisation assumes that processing and infrastructure capital costs have been sunk  Where key information is absent or deficient, Snowden has provided considered assumptions. Overall these assumptions are considered fair for the purpose of determining the Mineral Resource potential of the project.
Metallurgical factors or assumptions	 The basis for assumptions or predictions regarding metallurgical amenability. It is always necessary as part of the process of determining reasonable prospects for	 There has been a considerable amount of metallurgical test work to date.  All indications are that gold can be satisfactorily recovered from Sissingué ore using conventional extraction techniques.  Themetallurgical work is considered sufficient to determine

21 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
	eventual economic extraction to consider potential metallurgical methods, but the assumptions regarding metallurgical treatment processes and parameters made when reporting Mineral Resources may not always be rigorous. Where this is the case, this should be reported with an explanation of the basis of the metallurgical assumptions made.	that the Sissingué resource represents a deposit capable of economic extraction.  Recoveries used for the trial optimisation to determine the base of potentially economic material were based on studies carried out by Perseus as follows: o Oxide 92% o Transition 91% o Granite/Porphyry 90%  Sediments 78%
Environmental factors or assumptions	 Assumptions made regarding possible waste and process residue disposal options. It is always necessary as part of the process of determining reasonable prospects for eventual economic extraction to consider the potential environmental impacts of the mining and processing operation. While at this stage the determination of potential environmental impacts, particularly for a greenfields project, may not always be well advanced, the status of early consideration of these potential environmental impacts should be reported. Where these aspects have not been considered this should be reported with an explanation of the environmental assumptions made.	 There are other gold mines operating within Mali and Côte d’Ivoire in the general region where Sissingué is located.  The Sissingué Project area has no known environmental liabilities.  Perseus has been issued with an Environmental Permit to develop and operate an open pit mine and ore processing facility at Sissingué.
Bulk density	 Whether assumed or determined. If assumed, the basis for the assumptions. If determined, the method used, whether wet or dry, the frequency of the measurements, the nature, size and representativeness of the samples.  The bulk density for bulk material must have been measured by methods that adequately account for void spaces (vugs, porosity, etc.), moisture and differences between rock and alteration zones within the deposit.  Discuss assumptions for bulk density estimates used in the evaluation process of the different materials.	 A total of 770 bulk density measurements were taken from HQ and NQ drill core.  380 results are from oxide material, 132 from transitional material and 258 from fresh material.  Calculated means for the transitional (2.30 g/cm3) and fresh (2.73 g/cm3) weathering domains as well as the lateritic horizon (1.90 g/cm3) within the oxide domain were assigned to the block model.  The oxide domain, however, shows a gradient to higher densities with increasing depth and it is considered that a single value for the oxide domain would not be adequate. Instead, the bulk density for oxide blocks has been estimated using vertically orientated, inverse distance squared interpolation. The oxide bulk density ranges from 1.70 to 2.20 g/cm3.
Classification	 The basis for the classification of the Mineral Resources into	 The Sissingué Mineral Resource has been classified in the Measured, Indicated and Inferred categories, in accordance

22 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
	varying confidence categories.  Whether appropriate account has been taken of all relevant factors (ie relative confidence in tonnage/grade estimations, reliability of input data, confidence in continuity of geology and metal values, quality, quantity and distribution of the data).  Whether the result appropriately reflects the Competent Person’s view of the deposit.	with the 2012 Australasian Code for Reporting of Mineral Resources and Ore Reserves (JORC Code) and the CIM Definition Standards (CIM, 2005). A range of criteria has been considered in determining this classification including: - Geological continuity. - Data quality. - Drillhole spacing. - Estimation properties including kriging neighbourhood analysis to determine appropriate block size and search strategy. - Potential for economic extraction  The above parameters were used in combination to guide the manual digitising of strings on drill sections to control classification.  Trial optimisation has been run at a USD2,400 gold price (on the 2014 Mineral Resource) to define the base of potentially mineable material by open pit mining.  The Competent Person endorses the final results and classification.
Audits or reviews	 The results of any audits or reviews of Mineral Resource estimates.	 The Mineral Resource estimation procedure and results have been internally peer reviewed by Snowden.  During 2016 Jacqui Coombes of Coombes Capability carried out a review of the Sissingué diamond and RC drilling, and the potential impacts on the resource modelling processes. Jacqui made a series of recommendations which were considered in the 2016 update.
Discussion of relative accuracy/ confidence	 Where appropriate a statement of the relative accuracy and confidence level in the Mineral Resource estimate using an approach or procedure deemed appropriate by the Competent Person. For example, the application of statistical or geostatistical procedures to quantify the relative accuracy of the resource within stated confidence limits, or, if such an approach is not deemed appropriate, a qualitative discussion of the factors that could affect the relative accuracy and confidence of the estimate.  The statement should specify whether it relates to global or local estimates, and, if local, state the tonnages, which should be relevant to technical and economic evaluation. Documentation should include assumptions made and the procedures used.  These statements of relative accuracy and confidence of the estimate should be compared with production data, where available.	 No studies of relative accuracy have been carried out. The classification applied reflects the confidence in the Mineral Resource.

23 | P a g e

==> picture [120 x 37] intentionally omitted <==

1.4 Sissingué JORC Code 2012 Section 4 Estimation and Reporting of Ore Reserves

This section has been prepared by RPM to support the Statement Ore Reserves for Sissingué as of 31st March 2017.

Criteria	JORC Code explanation	Commentary
Mineral Resource estimate for conversion to Ore Reserves	 Description of the Mineral Resource estimate used as a basis for the conversion to an Ore Reserve.  Clear statement as to whether the Mineral Resources are reported additional to, or inclusive of, the Ore Reserves.	 The Mineral Resources for Sissingué were reported by Snowden Mining Industry Consultants (Snowden) in December 2016.  The Competent Person who prepared the Mineral Resource estimate is L. Olssen who is a chartered professional member of the Australasian Institute of Mining and Metallurgy  Mineral Resources quoted in this report are inclusive ofOreReserves
Site visits	 Comment on any site visits undertaken by the Competent Person and the outcome of those visits.  If no site visits have been undertaken indicate why this is the case.	 No site visit was undertaken specifically for this Ore Reserve Statement however an RPM employee at the time, Mr Paul Payne visited the Project site in August 2010.  As a RPM colleague had already been to site, and given it is a greenfields site, it was deemed that it would not be necessary that the Ore Reserves CP also go to site.
Study status	 The type and level of study undertaken to enable Mineral Resources to be converted to Ore Reserves.  The Code requires that a study to at least Pre-Feasibility Study level has been undertaken to convert Mineral Resources to Ore Reserves. Such studies will have been carried out and will have determined a mine plan that is technically achievable and economically viable, and that material Modifying Factors have been considered.	 The Mineral Resources have been converted to Ore Reserves by means of a Life of Mine plan including economic assessment.  Key aspects of the study were technically achievable pit designs based on Pit Limit Optimisation. These designs were also assessed to ensure economic viability.
Cut-off parameters	 The basis of the cut-off grade(s) or quality parameters applied.	 The cut-off grade is based on the economic parameters developed for the Operation. The cut-off grade varies due to different material types as follows Oxide/Laterite 0.45g/t Transitional 0.85g/t Granite/porphyry 0.85 g/t Sediment 1.05g/t
Mining factors or assumptions	 The method and assumptions used as reported in the Pre-Feasibility or Feasibility Study to convert the Mineral Resource to an Ore Reserve (i.e. either by application of appropriate factors by optimisation or by preliminary or detailed design).  The choice, nature and appropriateness of the selected mining method(s) and other mining parameters including associated design issues such as pre- strip, access, etc.  The assumptions made regarding geotechnical parameters (eg pit slopes, stope sizes, etc), grade control and pre- production drilling.  The major assumptions made and Mineral Resourcemodelusedforpit	 The proposed mining method is drill and blasting 5 m benches and selective mining by hydraulic excavators mining in 2.5 m flitches and loading standard off-highway rear dump trucks hauling to the ROM pad, surface ore stockpiles and waste disposal dumps. These would be supported by front-end loaders for stockpile rehandle.  The chosen method of mining is conventional open pit mining utilising hydraulic excavators and trucks, mining bench heights of 5 m in ore and 10m in waste.  The economic pit shell was defined using Whittle 4X pit optimisation software (“Whittle 4X”) with inputs such as geotechnical parameters, ore loss and dilution, metallurgical recovery and mining costs.  The pit optimisation was run with revenue generated only by Measured and Indicated Mineral Resources. No value was allocated to Inferred Mineral Resources.

24 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
	and stope optimisation (if appropriate).  The mining dilution factors used.  The mining recovery factors used.  Any minimum mining widths used.  The manner in which Inferred Mineral Resources are utilised in mining studies and the sensitivity of the outcome to their inclusion.  The infrastructure requirements of the selected mining methods.	 Whittle 4X input parameters were generally based on Perseus’s other operating site experience, supporting technical studies, and contractor estimates for mining costs.  Geotechnical parameters for Sissingué vary depending on the material type and Pit Sector The inter-ramp slope angles are between 38 to 53 degrees.  Appropriate mining modifying factors such as ore loss, dilution and design parameters were used to convert the Mineral Resource to an Ore Reserve  An SMU of 5m east x 5m west x 5m high was selected resulting in an approximately 9% ore loss and 15% dilution.  Minimum mining width of 40 m was generally applied to the pit designs.  As the mine is a green fields project all surface infrastructure is required to be constructed to enable the aforementioned mining method to be successfully implemented.  RPM has not identified or been informed of any physical constraints to mining within the lease area. No property, infrastructure or environmental issues are known to exist which may limit the extent of mining within the mining lease.
Metallurgical factors or assumptions	 The metallurgical process proposed and the appropriateness of that process to the style of mineralisation.  Whether the metallurgical process is well-tested technology or novel in nature.  The nature, amount and representativeness of metallurgical test work undertaken, the nature of the metallurgical domaining applied and the corresponding metallurgical recovery factors applied.  Any assumptions or allowances made for deleterious elements.  The existence of any bulk sample or pilot scale test work and the degree to which such samples are considered representative of the orebody as a whole.  For minerals that are defined by a specification, has the ore reserve estimation been based on the appropriate mineralogy to meet the specifications?	 The Sissingué processing plant uses crushing, grinding gravity recovery and cyanide leaching to extract gold. The plant has a forecast throughput capacity of 1.9 Mtpa on oxide ore and 1.0 Mtpa on fresh ore.  The technology used in the processing plant is well proven in many other operations globally.  The processing testwork is representative of the different material types throughout the Mining area.  No material deleterious material has been identified  A pilot scale metallurgical test work has been carried out to what Lycopodium considered representative of the orebody as a whole.  The process metallurgical recovery for gold varies by material type: Oxide/Laterite 91% Transitional 91% Granite/porphyry 90% Sediment 83% (average recovery as variable based on feed head grade)
Environment	 The status of studies of potential environmental impacts of the mining and processing operation. Details of waste rock characterisation and the consideration of potential sites, status of design options considered and, where applicable, the status of approvals for process residue storage and waste dumps should be reported.	 No environmental issues are known to exist which will prevent open pit mining to commence. Perseus appears to have sufficient space available for waste dumps to store the expected quantities of mine waste rock associated with the Sissingué open pit Ore Reserve. Based on testing to date no potentially acid forming material has been identified.  Sissingué has sufficient capacity in its purpose designed and approved tailings storage facility to meet the requirements generated from mining and processing quantities listed in for the Sissingué Ore Reserve.

25 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
Infrastructure	 The existence of appropriate infrastructure: availability of land for plant development, power, water, transportation (particularly for bulk commodities), labour, accommodation; or the ease with which the infrastructure can be provided, or accessed.	 Power supply will be via dedicated diesel generators.  Water supply will be largely from groundwater extracted from dedicated boreholes and supplemented with possible extraction from the nearby Bagoe River.  Access to site will be via an upgrade of the road from Tengrela.  A camp for 150 people will be established to accommodate non-local employees.  An airstrip will be established.  Workshops, offices, storage of reagents and laboratory will be established at the processing plant.
Costs	 The derivation of, or assumptions made, regarding projected capital costs in the study.  The methodology used to estimate operating costs.  Allowances made for the content of deleterious elements.  The derivation of assumptions made of metal or commodity price(s), for the principal minerals and co- products.  The source of exchange rates used in the study.  Derivation of transportation charges.  The basis for forecasting or source of treatment and refining charges, penalties for failure to meet specification, etc.  The allowances made for royalties payable, bothGovernment and private.	 The mining cost as based on a schedule of rates provided by a Perseus mining contractor. All other operating costs have been provided by Perseus and its Consultants.  Non deleterious materials have been identified and costed  Gold is the only metal considered in the Ore Reserves  All cost are in US$  A Bullion and Refining cost of US$1.02/oz was applied.  A royalty of 4.5% of the metal price was applied plus an additional royalty of US$0.80/oz.
Revenue factors	 The derivation of, or assumptions made regarding revenue factors including head grade, metal or commodity price(s) exchange rates, transportation and treatment charges, penalties, net smelter returns, etc.  The derivation of assumptions made of metal or commodity price(s), for the principal metals, minerals and co- products.	 A gold price of US$1,150/oz was provided by Perseus for mine planning and is considered to err on the conservative compared to current published metal price forecasts at US$1,200/oz.  Economic modelling by Perseus is at US$1,200/oz.
Market assessment	 The demand, supply and stock situation for the particular commodity, consumption trends and factors likely to affect supply and demand into the future.  A customer and competitor analysis along with the identification of likely market windows for the product.  Price and volume forecasts and the basis for these forecasts.  For industrial minerals the customer specification, testing and acceptance requirementsprior to a supplycontract.	 The demand for gold is considered in the gold price used.  It was considered that gold will be in demand and marketable for beyond the project life.  The processing forecast and mine life are based on life of mine plans.

26 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
Economic	 The inputs to the economic analysis to produce the net present value (NPV) in the study, the source and confidence of these economic inputs including estimated inflation, discount rate, etc.  NPV ranges and sensitivity to variations in the significant assumptions and inputs.	 A schedule and economic model has been completed by Perseus on a pre-tax basis using the inputs outlined in this Statement. The assessment used a discount rate of 10% which is considered appropriate by RPM. The NPV and sensitivity analysis outcomes indicated that the Project is economically viable. A schedule and economic model has been completed by Perseus on a pre-tax basis using the Ore Reserves published in this Statement. The inputs used are as per those stated in the relevant sections of this Statement. The assessment used a discount rate of 10% which is considered appropriate by RPM.Error! Reference source not found.
Social	 The status of agreements with key stakeholders and matters leading to social licence to operate.	 Perseus has established relevant agreements with local stakeholders.  The mine plan for the operation of the Sissingué and Sissingué open pits includes the use of skilled expatriate workers and locally sourced skilled workers.
Other	 To the extent relevant, the impact of the following on the project and/or on the estimation and classification of the Ore Reserves:  Any identified material naturally occurring risks.  The status of material legal agreements and marketing arrangements.  The status of governmental agreements and approvals critical to the viability of the project, such as mineral tenement status, and government and statutory approvals. There must be reasonable grounds to expect that all necessary Government approvals will be received within the timeframes anticipated in the Pre-Feasibility or Feasibility study. Highlight and discuss the materiality of any unresolved matter that is dependent on a third party on which extractionofthereserveis contingent.	 Error! Reference source not found.The estimate of Ore Reserves for the Sissingué Open Pits are not, to RPM’s knowledge, materially affected by any known environmental, permitting, legal, title, taxation, socio- economic, marketing, political or other relevant factors that would prevent the classification of Ore Reserves
Classification	 The basis for the classification of the Ore Reserves into varying confidence categories.  Whether the result appropriately reflects the Competent Person’s view of the deposit.  The proportion of Probable Ore Reserves that have been derived from Measured Mineral Resources (if any).	 Ore Reserves have been classified based on the underlying Mineral Resources classifications and the level of detail in the mine planning. The Mineral Resources were classified as Measured, Indicated and Inferred. The Ore Reserves, based only on the Measured and Indicated Resources, have been classified as Proven and Probable Ore Reserves, respectively  Ore Reserves have been classified based on the underlying Mineral Resources classifications and the level of detail in the mine planning. The Mineral Resources were classified as Measured, Indicated and Inferred. The Ore Reserves, based only on the Measured and Indicated Resources, have been classified as Proven and Probable Ore Reserves, respectively.  The Ore Reserve is classified as Proved and Probable in accordance with the JORC Code, corresponding to the Mineral Resource classifications of Measured and Indicated and taking into account other factorswhererelevant. The deposit’s geological

27 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
		model is well constrained. The Ore Reserve classification is considered appropriate given the nature of the deposit, the moderate grade variability, drilling density, structural complexity and mining history. Therefore it was deemed appropriate to use Measured Mineral Resources as a basis for Proven Reserves and Indicated Mineral Resources as a basis for Probable Reserves  No Inferred Mineral Resources were included in the Ore Reserve estimate
Audits or reviews	 The results of any audits or reviews of Ore Reserve estimates.	 RPM has completed an internal review of the Ore Reserve estimate.  The JORC Code provides guidelines which set out minimum standards, recommendations and guidelines for the Public Reporting of exploration results, Mineral Resources and Ore Reserves. Within the JORC Code is a “Checklist of Assessment and Reporting Criteria” (Table 1 – JORC Code). This checklist has been used as a systematic method to undertake a review of the underlying Study used to report in accordance with the JORC Code  A LOM Plan was prepared based on the ROM mineable quantity contained within the pit designs. RPM reviewed the LOM Plan prepared by Perseus and believes it is reasonable and practical and confirmed that it was suitable for estimation of Ore Reserves. An economic model was prepared by Perseus that confirmed the operation to be economically viable 
Discussion of relative accuracy/ confidence	 Where appropriate a statement of the relative accuracy and confidence level in the Ore Reserve estimate using an approach or procedure deemed appropriate by the Competent Person. For example, the application of statistical or geostatistical procedures to quantify the relative accuracy of the reserve within stated confidence limits, or, if such an approach is not deemed appropriate, a qualitative discussion of the factors which could affect the relative accuracy and confidence of the estimate.  The statement should specify whether it relates to global or local estimates, and, if local, state the relevant tonnages, which should be relevant to technical and economic evaluation. Documentation should include assumptions made and the procedures used.  Accuracy and confidence discussions should extend to specific discussions of any applied Modifying Factors that may have a material impact on Ore Reserve viability, or for which there are remaining areas of uncertainty at the current study stage.  It is recognised that this may not be possible or appropriate in all circumstances. These statements of	 The accuracy and confidence of the inputs are, as a minimum, of a pre-feasibility level (for the global open pit Ore Reserves).  The key factors that are likely to affect the accuracy and confidence in the Ore Reserves are: o Accuracy of the underlying Resource Block Models; o Changes in gold prices and sales agreements; o Changes in metallurgical recovery; and o Mining loss and dilution.  The Ore Reserve has utilised all parameters provided by site as made available.  The accuracy of the underlying Mineral Resources is defined by the Resource Category that the Mineral Resources are assigned to. Only the highest categories of Resource classification, Measured and Indicated, have been used as a basis for estimating Ore Reserves.

28 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
	relative accuracy and confidence of the estimate should be compared with productiondata,where available.

29 | P a g e

==> picture [120 x 37] intentionally omitted <==

2. Appendix 2: JORC Table 1 For Bélé

Section 1 to 3 of Table 1 has been prepared by Snowden and can be referenced as part of the 2017 Statement of Mineral Resources for Sissingué.

2.1 Bélé JORC Code 2012 Section 1 Sampling techniques and data This section has been prepared by Snowden to support the Statement of Mineral Resources for Bélé as of 20th February 2017.

Criteria	JORC Code explanation	Commentary
Sampling techniques	Nature and quality of sampling (e.g. cut channels, random chips, or specific specialised industry standard measurement tools appropriate to the minerals under investigation, such as down hole gamma sondes, or handheld XRF instruments, etc.). These examples should not be taken as limiting the broad meaning of sampling. Include reference to measures taken to ensure sample representivity and the appropriate calibration of any measurement tools or systems used Aspects of the determination of mineralisation that are Material to the Public Report. In cases where ‘industry standard’ work has been done this would be relatively simple (e.g. ‘reverse circulation drilling was used to obtain 1 m samples from which 3 kg was pulverised to produce a 30 g charge for fire assay’). In other cases, more explanation may be required, such as where there is coarse gold that has inherent sampling problems. Unusual commodities or mineralisation types (e.g. submarine nodules) may warrant disclosure of detailed information.	Samples at Bélé were collected using standard drilling techniques: reverse circulation (RC) and diamond drilling (DD). RC samples were collected in 1 m intervals at rig mounted cyclone. Samples from the first 13 RC holes were composited to 4 m length (111 samples, <1% of all RC samples). All other RC drilling was composited to 2 m samples, by equal weight. Sample weight was nominally 3 kg for composited samples. Diamond core was generally sampled at 1 m intervals. The first 13 RC holes were assayed by 1 kg 24-hour bottle roll, all subsequent DD and RC samples by 50 g or 40 g fire assay. All analytical work up till March 2016 was carried out by independent, commercial laboratory Bureau Veritas Minerals Laboratory (BMVL) in Abidjan, Côte d’Ivoire. Analytical work for the recent drill holes program between November 2016 and January 2017, was carried out by independent, commercial laboratory Actlabs Burkina Faso SARL (Actlabs) in Ouagadougou, Burkina Faso.
Drilling techniques	Drill type (e.g. core, reverse circulation, open-hole hammer, rotary air blast, auger, Bangka, sonic, etc.) and details (e.g. core diameter, triple or standard tube, depth of diamond tails, face- sampling bit or other type, whether core is oriented and if so, by what method, etc.).	All RC drilling used a 5.25 inch diameter face sample bit. Diamond core drilling was carried out using HQ3 in the oxide and transitional rock and NQ2 in fresh rock. Diamond core was generally oriented using a spear.

30 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
Drill sample recovery	Method of recording and assessing core and chip sample recoveries and results assessed. Measures taken to maximise sample recovery and ensure representative nature of the samples. Whether a relationship exists between sample recovery and grade and whether sample bias may have occurred due to preferential loss/gain of fine/coarse material.	RC samples were weighed at 1 m intervals and recoveries back- calculated using nominal hole diameter and expected density values. Recoveries average between 60% and 75% in strongly weathered material depending on rock type, around 75% in the transition zone and >85% in fresh rock. Recovered length of diamond samples were measured. The overall recovery of 93% is considered good, although Snowden notes that the recovery is lower in the oxide and transitional materials. No apparent relationship exists between sample recovery and grade for diamond drilling.
Logging	Whether core and chip samples have been geologically and geotechnically logged to a level of detail to support appropriate Mineral Resource estimation, mining studies and metallurgical studies. Whether logging is qualitative or quantitative in nature. Core (or costean, channel, etc.) photography. The total length and percentage of the relevant intersections logged.	RC drill chips were logged geologically, including rock type, weathering, alteration type and intensity (where recognisable), vein quartz content in estimated percentage, sulphide minerals and estimated content. Diamond drill core was geologically and structurally logged. Geological logging is identical to RC logging. Structural logging includes joints, fractures, roughness and infill type of structures and veins as well as recovery and RQD. Logging was qualitative (descriptive) and semi-quantitative (estimates) in nature. All diamond core was photographed in the core boxes. RC drill chips were glued on chip boards for visual reference for each hole. All drill holes (RC and diamond) were logged in full.

31 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
Subsampling techniques and sample preparation	If core, whether cut or sawn and whether quarter, half or all core taken. If non-core, whether riffled, tube sampled, rotary split, etc. and whether sampled wet or dry. For all sample types, the nature, quality and appropriateness of the sample preparation technique. Quality control procedures adopted for all subsampling stages to maximise representivity of samples. Measures taken to ensure that the sampling is representative of the in situ material collected, including for instance results for field duplicate/second-half sampling. Whether sample sizes are appropriate to the grain size of the material being sampled.	Diamond core was cut in half using a diamond saw. The “right” side of the core was submitted for analysis, the other half stored in core trays. All RC samples were collected at the drill site at 1 m interval and split using a multi-stage riffle splitter. Each two (resp. four) consecutive samples were composited in one bag by equal weight. 621 or 4% of RC samples were logged as wet. Wet samples were sub-sampled using a spear. 888 or 5% of RC samples were logged as damp. Sample preparation of diamond core and RC chips used industry standard techniques. After drying, the entire sample was subject to a primary crush, 200 g of sub-sample was collected and pulverised. Internal laboratory checks required at least 90% of the pulp passing -75 microns. Laboratory QAQC includes the use of internal standards, certified reference materials and pulp replicates. Field sampling QAQC procedures included the use of certified reference materials inserted at a rate of 1 in 20. Between one and two field duplicates were taken for each RC hole, preferably within mineralised intervals. The results of duplicate analysis show no bias, but only moderate repeatability. Field duplicates of diamond core were not taken as ¼ core is considered inappropriate for comparison. Coarse crush and pulp duplicates were taken for RC and diamond samples during the recent drilling program and show good precision. Some RC samples at depth were identified as having downhole contamination and resultant smearing of grades as a result of wet drilling in ‘sticky’ material, with the samples being ‘hung up’ in the cyclone and subsequently contaminating later samples. This issue appears to only occur in a few drill holes and is not as prevalent as what was seen at the nearby Sissingué deposit. As a result of this, all RC holes in the pit area were reviewed and any suspected of containing smeared assays were removed from the dataset prior to estimation. 277 m from four RC drill holes (<1% of the samples) were removed due to suspected downhole contamination. In addition, 2016 drilling focused on diamond drill holes to confirm areas with RC drilling in the core of the deposit. With the exception of the issue noted above, the sub-sampling is considered appropriate and representative.

32 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
Quality of assay data and laboratory tests	The nature, quality and appropriateness of the assaying and laboratory procedures used and whether the technique is considered partial or total. For geophysical tools, spectrometers, handheld XRF instruments, etc., the parameters used in determining the analysis including instrument make and model, reading times, calibrations factors applied and their derivation, etc. Nature of quality control procedures adopted (e.g. standards, blanks, duplicates, external laboratory checks) and whether acceptable levels of accuracy (i.e. lack of bias) and precision have been established.	A small number of samples from initial holes were assayed by bottle roll with AAS finish, all subsequent RC and DD samples were assayed by standard 50 g fire assay with AAS finish for total gold content. No geophysical tools were used to determine any element concentrations. QAQC procedures included One to two field duplicates per RC hole Certified blanks inserted at one in 40 Certified standards at one in 20 Internal laboratory standards, duplicates and repeats. Review of the standards results indicates that Actlabs tends to undercall the gold standards for low grade samples by around 5% to 10%. As a result, umpire analysis was carried out on two batches using BMVL. The umpire results show that BMVL reports the low grade standards accurately. BMVL reports around a 5% to 10% higher gold grade for the low grade samples between 0.3 and 0.8 g/t gold. Results are comparable at all other grade ranges. As a result of the above analysis, Snowden considers the Actlabs results acceptable for resource estimation, with the acknowledgement that the low grade samples are slightly conservative. Given the Mineral Resource reporting cut-off of 0.8 g/t gold, this should not have a material impact on the Mineral Resource. With the exception of the item above, the QAQC shows acceptable precision and no bias. Overall assaying quality is considered adequate.
Verification of sampling and assaying	The verification of significant intersections by either independent or alternative company personnel. The use of twinned holes. Documentation of primary data, data entry procedures, data verification, data storage (physical and electronic) protocols. Discuss any adjustment to assay data.	Significant intersections are not highlighted; all RC cuttings and diamond core are systematically sampled. As part of the 2016 drilling program, several RC drill holes with suspected downhole smearing due to contamination at Bélé East were twinned with diamond drill holes. As a result, the suspect RC drill holes were removed as discussed previously. Drill holes information for both RC and diamond core holes is captured at the drill site on paper. All hard copies are handed over to the database assistant at the site office and the information provided on paper is then entered into a database. All hard copies are kept at the Tengrela site office. Downhole survey data and collar survey data are provided by the drilling contractors and surveyors respectively in digital format. Perseus maintains a centralised database for its various operations in Ghana and Côte d’Ivoire. Database administration is based in Perseus’s office in Accra/Ghana and under the supervision of the company’s Resource Geologist. 2,082 below detection limit samples were reset to 0.001 g/t gold. No other adjustments were made to the raw assay data with the exception of the removal of any RC samples with suspected smearing of grades as previously discussed. Top cutting is only applied after database compositing and statistical analysis and prior to resource estimation.

33 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
Location of data points	Accuracy and quality of surveys used to locate drill holes (collar and downhole surveys), trenches, mine workings and other locations used in Mineral Resource estimation. Specification of the grid system used. Quality and adequacy of topographic control.	Most RC and diamond drill holes collars were surveyed by the company’s surveyor in 2015 and 2016 using differential GPS equipment. 22 holes, including two holes with diamond tails, could not be found at the time and were not surveyed. In these cases, the original coordinates taken by handheld GPS were used. On average, the difference between handheld and differential GPS is less than 2 m in the X and Y directions. Downhole surveys were conducted by the drill contractors using a FlexIT tool at 30 m intervals for the diamond drill holes. Generally RC holes only have the collar azimuth and inclination measured. The WGS84 UTM Zone 29 North grid system is used. The topography covering the extent of the Resource model was created as a digital terrain model (DTM) in Surpac using the surveyed drill hole collar data and an additional 77,767 points established at 3 m intervals by differential GPS during 100 m spaced geophysical traverses.
Data spacing and distribution	Data spacing for reporting of Exploration Results Whether the data spacing and distribution is sufficient to establish the degree of geological and grade continuity appropriate for the Mineral Resource and Ore Reserve estimation procedure(s) and classifications applied. Whether sample compositing has been applied.	Nominal drill holes spacing over the resource area is predominantly 20 m by 40 m to 40 m by 40 m at Bélé West and 20 m by 20 m at Bélé East. Data spacing is sufficient to establish grade and geological continuity appropriate to the resource estimation procedures and classifications applied. Diamond and RC samples within the resource have been composited to 2 m.
Orientation of data in relation to geological structure	Whether the orientation of sampling achieves unbiased sampling of possible structures and the extent to which this is known, considering the deposit type. If the relationship between the drilling orientation and the orientation of key mineralised structures is considered to have introduced a sampling bias, this should be assessed and reported if material.	Orientation of drill holes is approximately perpendicular to the strike of the geology and mineralisation at Bélé West. At Bélé East, drill holes are angled to cross the steep dip of the geological domains. At Bélé East, 12 early RC holes have been drilled along exploration fences oriented towards the east and hence sub-parallel to the mineralisation. Three of these holes intercepted significant mineralisation. These intercepts have been verified by holes drilled in the opposite direction, however, they have been removed from the database for estimation to ensure no bias occurs due to the orientation.
Sample security	The measures taken to ensure sample security.	Chain of custody is managed by Perseus. Samples are stored on site until conveyed to the Bureau Veritas laboratory in Abidjan. Once dispatched, Perseus personnel have no further involvement in the preparation or analysis of the samples.

34 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
Audits or reviews	The results of any audits or reviews of sampling techniques and data.	Several reviews of sampling techniques were carried out by the company’s senior personnel during site visits, with acceptable conclusions. Basic drill data validation has been carried out by Snowden during the preparation of the Mineral Resource estimate. No material issues were noted with the exception of the down-dip and potentially smeared RC samples mentioned previously. In addition Snowden reviewed the drilling and sampling procedures for diamond and RC drilling at the Bélé deposit. Snowden also visited the core shed and reviewed the diamond core sampling processes and diamond core, RC chip boards and logging procedures. No material issues in the procedures were noted. For the 2016 drilling, where RC drilling is used, the RC drilling is conducted until wet samples are returned, at which point the drilling is changed to diamond drilling to avoid the contamination of RC samples.

35 | P a g e

==> picture [120 x 37] intentionally omitted <==

2.2 Bélé JORC Code 2012 Section 2 Reporting of Exploration Results

This section has been prepared by Snowden to support the Statement of Mineral Resources for Bélé as of 20[th] February 2017.

Criteria	JORC Code explanation	Commentary
Mineral tenement and land tenure status	Type, reference name/number, location and ownership including agreements or material issues with third parties such as joint ventures, partnerships, overriding royalties, native title interests, historical sites, wilderness or national park and environmental settings. The security of the tenure held at the time of reporting along with any known impediments to obtaining a licence to operate in the area.	Perseus, through its 100% owned subsidiary Occidental Gold SARL holds an Exploration Licence Permit PR259, issued by the Ivorian government in December 2012, valid for mineral exploration (Decree “2012-1172”). The tenement is fully owned by Perseus, through its subsidiary, with the Ivorian government holding a statutory 10% free carried interest. The Exploration Licence expired on 19 December 2015 and is currently under application for renewal. The application has been approved by the General Director of Mines and Geology and is awaiting final signature of the Minister of Mines and Industry. The initial licence covered an area of 398 km². The renewed licence has been reduced to an area of 298.5 km². There are no known impediments with respect to exploration or mining.
Exploration done by other parties	Acknowledgment and appraisal of exploration by other parties.	Perseus is not aware of any previous exploration activities.
Geology	Deposit type, geological setting and style of mineralisation.	The Bélé gold deposits are located within a north- westerly striking splay of the Syama-Boundiali Greenstone Belt. At Bélé, Birimian aged rocks comprise a sequence of metasedimentary rocks and subordinate mafic volcanics that have been intruded by a nearly circular granitoid body approximately 4 km in diameter. The sequence has also been intruded by numerous felsic dykes of various compositions. Gold mineralisation at both Bélé East and Bélé West is associated with deformation zones developed at and adjacent to the margins of the granitoid intrusion. Gold is associated with disseminated pyrite and lesser pyrrhotite hosted by both mafic and felsic lithologies where they feature chlorite-sericite-calcite alteration. Vein-hosted mineralisation is rare. Bélé West mineralisation is interpreted to extend around 1 km in strike, 50 m thickness (comprising several lodes up to 20 m thick each) and to a depth of 150 m. Bélé East mineralisation extends around 500 m along strike, 130 m thickness (comprising several lodes up to 20 m thick each) and to a depth of 170 m. The currently defined mineralisation in both areas is open at depth but appears to be closed out along strike.

36 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
Drill holes Information	A summary of all information material to the understanding of the exploration results including a tabulation of the following information for all Material drill holes: easting and northing of the drill hole collar elevation or RL (Reduced Level – elevation above sea level in metres) of the drill hole collar dip and azimuth of the hole downhole length and interception depth hole length. If the exclusion of this information is justified on the basis that the information is not Material and this exclusion does not detract from the understanding of the report, the Competent Person should clearly explain why this is the case.	The Bélé drill holes data includes RC, diamond and aircore drill holes. Aircore drill holes were used as a guide to interpretation but were not used for estimation due to the poor quality of aircore samples. In additional 376 m (from four drill holes; 1% of the samples) of RC drilled down-dip to the mineralisation and 277 m (from four drill holes; <1% of the samples) of RC with suspected downhole smearing due to contamination were excluded in Bélé East. Drilling used for the Mineral Resource includes: 274 RC drill holes for 21,937 m 47 RC drill holes with diamond tail for 5,431 m 54 diamond drill holes for 2,599 m. Drill holes were set up using a compass. Compass readings and all downhole survey azimuth readings were adjusted according to the magnetic deviation. Exploration results have been reported previously. All drilling is included for Mineral Resource estimation.
Data aggregation methods	In reporting Exploration Results, weighting averaging techniques, maximum and/or minimum grade truncations (e.g. cutting of high grades) and cut-off grades are usually Material and should be stated. Where aggregate intercepts incorporate short lengths of high grade results and longer lengths of low grade results, the procedure used for such aggregation should be stated and some typical examples of such aggregations should be shown in detail. The assumptions used for any reporting of metal equivalent values should be clearly stated.	NA; exploration results are not reported.
Relationship between mineralisation widths and intercept lengths	These relationships are particularly important in the reporting of Exploration Results. If the geometry of the mineralisation with respect to the drill hole angle is known, its nature should be reported. If it is not known and only the down hole lengths are reported, there should be a clear statement to this effect (e.g. ‘downhole length, true width not known’).	The majority of drilling is perpendicular to the mineralisation. Some scissor holes have also been drilled, particularly in Bélé East where the mineralisation is sub-vertical. Three of these holes intercepted significant mineralisation. These intercepts have been verified by holes drilled in the opposite direction, however, they have been removed from the database for estimation to ensure no bias occurs due to the orientation. Areas of the main northeast trending limb of Bélé West are not optimally orientated as they are drilled east-west and the mineralisation is trending to the northeast. The orientation is acceptable to define the limits of the mineralisation however.
Diagrams	Appropriate maps and sections (with scales) and tabulations of intercepts should be included for any significant discovery being reported These should include, but not be limited to a plan view of drill hole collar locations and appropriate sectional views.	Location plans and example sections are included in the Mineral Resource technical documentation.

37 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
Balanced reporting	Where comprehensive reporting of all Exploration Results is not practicable, representative reporting of both low and high grades and/or widths should be practiced to avoid misleading reporting of Exploration Results.	NA; all drilling is included in the Mineral Resource.
Other substantive exploration data	Other exploration data, if meaningful and material, should be reported including (but not limited to): geological observations; geophysical survey results; geochemical survey results; bulk samples – size and method of treatment; metallurgical test results; bulk density, groundwater, geotechnical and rock characteristics; potential deleterious or contaminating substances.	NA; all drilling is included in the Mineral Resource.
Further work	The nature and scale of planned further work (e.g. tests for lateral extensions or depth extensions or large-scale step-out drilling). Diagrams clearly highlighting the areas of possible extensions, including the main geological interpretations and future drilling areas, provided this information is not commercially sensitive.	Further resource definition drilling depends upon preliminary economic evaluation of the resources defined to date. Further exploration is proposed for other areas of the exploration permit.

38 | P a g e

==> picture [120 x 37] intentionally omitted <==

2.3 Bélé JORC Code 2012 Section 3 Estimation and Reporting of Mineral Resources This section has been prepared by Snowden to support the Statement of Mineral Resources for Bélé as of 20[th] February 2017.

Criteria	JORC Code explanation	Commentary
Database integrity	Measures taken to ensure that data has not been corrupted by, for example, transcription or keying errors, between its initial collection and its use for Mineral Resource estimation purposes. Data validation procedures used.	Perseus validates the data routinely during data entry and data import into its database platform, Maxwell’s Datashed software, and during importation into Surpac. This includes checks for Duplicate or erroneous collars information Missing samples Down-hole from-to interval inconsistencies Overlapping samples Records beyond hole depth Missing assays Invalid geological codes Import of digitally communicated assay results into Datashed is automated and does not necessitate manual interference (such as copy-and-paste commands). Snowden carried out basic validation checks as part of preparing the data for estimation. The database checks undertaken by Snowden are listed below; no significant issues were identified: The downhole sampling intervals are consistent with no overlapping sample intervals. Assay values are within realistic limits. 2,082 below detection limit samples were reset to 0.001 g/t gold. The sample tables were checked to ensure there are no duplicate sample records. There are no missing or incomplete collar survey coordinates. In addition, Snowden reviewed all RC samples within the pit area to assess them for potential downhole contamination and resultant smearing. Any samples suspected of containing smeared assays were removed from the dataset prior to estimation. This impacts 277 m from four RC drill holes (<1% of the samples).
Site visits	Comment on any site visits undertaken by the Competent Person and the outcome of those visits. If no site visits have been undertaken indicate why this is the case.	Steffen Brammer (Perseus) has visited the site several times and on a regular basis since the commencement of Perseus’ exploration activities and during various drilling campaigns. Lynn Olssen (Senior Principal Consultant) of Snowden Mining Industry Consultants visited site during December 2016. Lynn reviewed the drilling and sampling procedures for diamond and RC drilling at the Bélé deposit. Lynn also visited the core shed and reviewed the diamond core sampling processes and diamond core, RC chip boards and logging procedures. No material issues in the procedures were noted.

39 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
Geological interpretation	Confidence in (or conversely, the uncertainty of) the geological interpretation of the mineral deposit. Nature of the data used and of any assumptions made. The effect, if any, of alternative interpretations on Mineral Resource estimation. The use of geology in guiding and controlling Mineral Resource estimation. The factors affecting continuity both of grade and geology.	Mineralisation domains, weathering and lithology were originally interpreted by Perseus. Snowden reviewed and updated this interpretation for the March 2016 update. Mineralisation at Bélé occurs as a series of lodes subparallel to the mafic lithologies. The lodes cross the mafics-granite boundary in places. In Bélé East there is a wider dilation of mineralisation in the central area which appears to be associated with a kink in the lithology, and is likely structurally related. The mineralisation domains were defined based on a nominal 0.2 g/t gold to 0.5 g/t gold cut-off together with the geological logging and lithology interpretation. The cut-off used is observed as a population change in the global log- probability plot. Note, the drill holes data includes RC, diamond and aircore drill holes. Aircore drill holes were used as a guide to interpretation but were not used for estimation due to the poor quality of aircore samples. The mineralisation domains were used as hard boundaries to control estimation. Snowden considers that any alternative interpretation would not have a material effect on the Mineral Resource estimate.
Dimensions	The extent and variability of the Mineral Resource expressed as length (along strike or otherwise), plan width, and depth below surface to the upper and lower limits of the Mineral Resource.	Bélé West mineralisation is interpreted to extend around 1 km in strike, 140 m across strike and to a depth of 150 m. Bélé East mineralisation extends around 500 m along strike, 150 m across strike and to a depth of 170 m. The currently defined mineralisation in both areas is open at depth but appears to be closed out along strike.

40 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
Estimation and modelling techniques	The nature and appropriateness of the estimation technique(s) applied and key assumptions, including treatment of extreme grade values, domaining, interpolation parameters and maximum distance of extrapolation from data points. If a computer assisted estimation method was chosen include a description of computer software and parameters used. The availability of check estimates, previous estimates and/or mine production records and whether the Mineral Resource estimate takes appropriate account of such data. The assumptions made regarding recovery of by-products. Estimation of deleterious elements or other non-grade variables of economic significance (e.g. sulphur for acid mine drainage characterisation). In the case of block model interpolation, the block size in relation to the average sample spacing and the search employed. Any assumptions behind modelling of selective mining units. Any assumptions about correlation between variables. Description of how the geological interpretation was used to control the resource estimates. Discussion of basis for using or not using grade cutting or capping. The process of validation, the checking process used, the comparison of model data to drill hole data, and use of reconciliation data if available.	Snowden estimated gold grades using ordinary block kriging (parent cell estimates) using Datamine software. The grades were estimated using the mineralisation domains for each area as hard boundaries. Ordinary kriging was selected as appropriate for estimation on the basis that coefficients of variation are generally low to moderate within the deposit, with top cuts used where required to control the influence of extreme grades. Discretisation of 4 by 4 by 4 was applied during estimation. Parent block size was based on kriging neighbourhood analysis (KNA) results and the geometry of the two areas; a 10 mE by 10 mN by 5 mRL parent block size was selected for both areas. The average drill holes spacing is 20 m by 20 m to 20 m by 40 m. Review of the histograms and log-probability plots shows that there is a high coefficient of variation (CV) and outliers evident within the Bélé East area. As a result, a top cut of 30 g/t Au was applied to the data in this area prior to estimation. The Bélé West area has a maximum value of <30 g/t gold and does not appear to require top cutting. Dynamic anisotropy was used for estimation, whereby the local dip and azimuth of the mineralised lodes was estimated into each block in the model and the search and variogram orientations were locally adjusted to reflect the geological orientation. This method allows the estimate to better reflect the changing orientation and undulating nature of the lodes. Search parameters were defined based on the ranges of continuity seen in the variograms and KNA results. Estimates were run using a minimum of 6 and a maximum of 14 informing composites for the first search pass. The second search pass used a minimum of six and a maximum of eight informing composites, with the search ellipse doubled in size to inform blocks not informed during the first search pass. A third search pass with the number of samples reduced to a minimum of four and a maximum of eight, with six times the search ellipse size, was to used where an estimate could not be obtained using the second search pass. To ensure that data from multiple drill holes were used during the block estimation, a maximum of three composites were allowed from each drill holes. Final grade estimates were validated by: undertaking global grade comparisons with the input drill holes composites; visual validation of block model cross sections; and by grade trend plots. In addition, a theoretical global change of support was carried out to validate the level of smoothing in the estimate. The resultant estimate contains less tonnes at a higher grade within the main domains compared to the previous estimate. This is a result of the additional data allowing for a more locally accurate estimate. Application of the constraining pit shell has removed the deeper portions of the Inferred Resource, together with some blocks at the northern and southern extents. This has resulted in an overall reduction in tonnes and ounces compared to the previous Mineral Resource.
Moisture	Whether the tonnages are estimated on a dry basis or with natural moisture, and the method of determination of the moisture content.	All tonnages are calculated and reported on a dry tonnes basis.

41 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
Cut-off parameters	The basis of the adopted cut-off grade(s) or quality parameters applied.	The Mineral Resource has been reported by resource classification and weathering above a 0.8 g/t gold cut-off. The cut-off grade is based on preliminary engineering work which indicates a 0.75 g/t Gold to 0.85 g/t gold cut-off will be applicable for mining, depending on the degree of weathering.
Mining factors or assumptions	Assumptions made regarding possible mining methods, minimum mining dimensions and internal (or, if applicable, external) mining dilution. It is always necessary as part of the process of determining reasonable prospects for eventual economic extraction to consider potential mining methods, but the assumptions made regarding mining methods and parameters when estimating Mineral Resources may not always be rigorous. Where this is the case, this should be reported with an explanation of the basis of the mining assumptions made.	Mining is assumed to be traditional drill and blast open cut mining.
Metallurgical factors or assumptions	The basis for assumptions or predictions regarding metallurgical amenability. It is always necessary as part of the process of determining reasonable prospects for eventual economic extraction to consider potential metallurgical methods, but the assumptions regarding metallurgical treatment processes and parameters made when reporting Mineral Resources may not always be rigorous. Where this is the case, this should be reported with an explanation of the basis of the metallurgical assumptions made.	Metallurgical work has been initiated and is currently ongoing. Preliminary results suggest that the ore is amenable to the treatment processes considered for the company’s nearby Sissingué project.
Environmental factors or assumptions	Assumptions made regarding possible waste and process residue disposal options. It is always necessary as part of the process of determining reasonable prospects for eventual economic extraction to consider the potential environmental impacts of the mining and processing operation. While at this stage the determination of potential environmental impacts, particularly for a greenfields project, may not always be well advanced, the status of early consideration of these potential environmental impacts should be reported. Where these aspects have not been considered this should be reported with an explanation of the environmental assumptions made.	There are other gold mines operating within Mali and Côte d’Ivoire in the general region where Bélé is located. There are no known environmental impediments to mining.

42 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
Bulk density	Whether assumed or determined. If assumed, the basis for the assumptions. If determined, the method used, whether wet or dry, the frequency of the measurements, the nature, size and representativeness of the samples. The bulk density for bulk material must have been measured by methods that adequately account for void spaces (vugs, porosity, etc.), moisture and differences between rock and alteration zones within the deposit. Discuss assumptions for bulk density estimates used in the evaluation process of the different materials.	Bulk density values were determined by measurements on available core and RC chips. The values assigned to the models are based on the weathering and lithology domains and range from 1.60 t/m3in the oxide granites to 2.76 t/m3in the fresh mafic volcanics. Bulk density of fresh material was determined by 93 laboratory measurements of half core pieces (10 cm pieces) using an Archimedes method of weight in water and weight in air. Full core trays were also weighed and bulk density calculated by measuring the length of the core within the tray and calculating the volume. This method compares well to the laboratory measurements for the granites but resulted in an unrealistically low bulk density for the mafic volcanics. Bulk density for the transitional material was assigned based on 98 whole core tray measurements. In-house Archimedes style measurements were also taken on whole core pieces (10 cm pieces), however, while the granite values are the same for both methods, the mafic volcanics appear unrealistically low (lower than the granite) for the in-house measurements. Bulk density for the oxide was assigned based on 47 in- house measurements which appear reasonable. While core tray measurements were also calculated, these are considered prone to bias in the oxides due to the poorer recovery in the oxide material. The core tray values are higher than the in-house measurements. Previous transitional and oxide bulk densities were based on weighing RC sample bags which is acknowledged to be a suboptimal method. Snowden considers that the revised bulk density values (which are typically lower than those used previously) are more robust.
Classification	The basis for the classification of the Mineral Resources into varying confidence categories. Whether appropriate account has been taken of all relevant factors (i.e. relative confidence in tonnage/grade estimations, reliability of input data, confidence in continuity of geology and metal values, quality, quantity and distribution of the data). Whether the result appropriately reflects the Competent Person’s view of the deposit.	The resource estimate has been classified as an Indicated and Inferred Resource in accordance with the 2012 Australasian Code for Reporting of Mineral Resources and Ore Reserves (JORC Code) and the CIM Definition Standards (CIM, 2005). A range of criteria has been considered in determining this classification including:  Geological continuity.  Data quality.  Drill holes spacing.  Estimation properties including kriging neighbourhood analysis to determine appropriate block size and search strategy.  Potential for economic extraction The above parameters were used in combination to guide the manual digitising of strings on drill sections to control classification. Typically Indicated Resources are defined in areas of 20 m by 20 m drilling at Bélé East and 40 m by 40 m drilling at Bélé West which shows more continuity in grade. Trial optimisation has been run at a US$2,400 gold price to define the base of potentially mineable material by open pit mining. The Competent Person endorses the final results and classification.

43 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
Audits or reviews	The results of any audits or reviews of Mineral Resource estimates.	The Mineral Resource has been internally reviewed by Snowden.
Discussion of relative accuracy/ confidence	Where appropriate a statement of the relative accuracy and confidence level in the Mineral Resource estimate using an approach or procedure deemed appropriate by the Competent Person. For example, the application of statistical or geostatistical procedures to quantify the relative accuracy of the resource within stated confidence limits, or, if such an approach is not deemed appropriate, a qualitative discussion of the factors that could affect the relative accuracy and confidence of the estimate. The statement should specify whether it relates to global or local estimates, and, if local, state the relevant tonnages, which should be relevant to technical and economic evaluation. Documentation should include assumptions made and the procedures used. These statements of relative accuracy and confidence of the estimate should be compared with production data, where available.	No studies of relative accuracy have been carried out. The classification applied reflects the confidence in the Mineral Resource. No production data is available.

2.4 Bélé JORC Code 2012 Section 4 Estimation and Reporting of Mineral Reserves

This section has been prepared by Perseus to support the Statement of Ore Reserves for Bélé as of 31[st] March 2017

Criteria	JORC Code explanation	Commentary
Mineral Resource estimate for conversion to Ore Reserves	 Description of the Mineral Resource estimate used as a basis for the conversion to an Ore Reserve.  Clear statement as to whether the Mineral Resources are reported additional to, or inclusive of, the Ore Reserves.	 The Mineral Resources for Bélé were reported by Snowden Mining Industry Consultants (Snowden) in February 2017  Error! Reference source not found.  Error! Reference source not found.
Site visits	 Comment on any site visits undertaken by the Competent Person and the outcome of those visits.  If no site visits have been undertaken indicatewhy thisis the case.	 Mr Paul Thompson, Perseus employee, visited the Project site in June 2015 and is a Fellow of the Australian Institute of Mining and Metallurgy.
Study status	 The type and level of study undertaken to enable Mineral Resources to be converted to Ore Reserves.  The Code requires that a study to at least Pre-Feasibility Study level has been undertaken to convert Mineral Resources to Ore Reserves. Such studies will have been carried out and will have determined a mine plan that is technically achievable and economically viable, and that material Modifying Factors have been considered.	 The Mineral Resources have been converted to Ore Reserves by means of a Life of Mine plan including economic assessment.  Key aspects of the study were technically achievable pit designs based on Pit Optimisation. These designs were also assessed to ensure economic viability.
Cut-off	 The basis ofthe cut-offgrade(s) or	 The cut-offgradeis based onthe economic

44 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
parameters	quality parameters applied.	parameters developed for the Operation. The cut-off grade varies due to different material types as follows Oxide – 0.65g/t Transitional – 0.95g/t Granite – 1.05g/t Mafic –1.20g/t
Mining factors or assumptions	 The method and assumptions used as reported in the Pre-Feasibility or Feasibility Study to convert the Mineral Resource to an Ore Reserve (i.e. either by application of appropriate factors by optimisation or by preliminary or detailed design).  The choice, nature and appropriateness of the selected mining method(s) and other mining parameters including associated design issues such as pre- strip, access, etc.  The assumptions made regarding geotechnical parameters (eg pit slopes, stope sizes, etc), grade control and pre- production drilling.  The major assumptions made and Mineral Resource model used for pit and stope optimisation (if appropriate).  The mining dilution factors used.  The mining recovery factors used.  Any minimum mining widths used.  The manner in which Inferred Mineral Resources are utilised in mining studies and the sensitivity of the outcome to their inclusion.  The infrastructure requirements of the selected mining methods.	 The chosen method of mining is conventional open pit mining utilising hydraulic excavators and trucks, mining 5 m bench heights, and utilising a number of flitches to minimise ore loss and waste rock dilution.  The economic pit shell was defined using Whittle 4X pit optimisation software (“Whittle 4X”) with inputs such as geotechnical parameters, ore loss and dilution, metallurgical recovery and economic parameters.  The pit optimisation was run with revenue generated only by Measured and Indicated Mineral Resources. No value was allocated to Inferred Mineral Resources.  Whittle 4X input parameters were generally based on regional operating experience and supporting technical studies.  Geotechnical parameters for Bélé vary depending on the material type and pit sector. The inter-ramp slope angles vary between 35 and 52 degrees.  Appropriate mining modifying factors such as ore loss, dilution, economics and design parameters were used to convert the Mineral Resource to an Ore Reserve  The mining dilution and recovery factors used for Bélé are 2% (East) and 4% (West) loss with 5% (East) and 10% (West) dilution.  Minimum mining width of 20 m was generally applied to the pit designs for the selected type of equipment.  As the mine is a greenfield project surface infrastructure is required to enable the aforementioned mining method to be successfully implemented.  The Bélé West and East pit footprints cut through creek lines. Therefore the pits will require creek diversions to be constructed before mining commences.
Metallurgical factors or assumptions	 The metallurgical process proposed and the appropriateness of that process to the style of mineralisation.  Whether the metallurgical process is well-tested technology or novel in nature.  The nature, amount and representativeness of metallurgical test work undertaken, the nature of the metallurgical domaining applied and the corresponding metallurgical recovery factors applied.  Any assumptions or allowances made for deleterious elements.  The existence of any bulk sample or pilot scale test work and the degree to which such samples are considered representative of the orebody as a whole.	 The Bélé ore material will be hauled to and processed at the Sissingué processing plant utilising the crushing, grinding and CIL circuit that is under construction at the time of compiling this report. The Sissingué processing plant has a capacity up to 1.5 Mtpa depending on the material type.  The Bélé ore will require a finer grind to 75μm compared to the Sissingué ore (106μm). Bélé ore will be either batch treated through the Sissingué processing plant and/or processed after completion of processing of all Sissingué ore  The technology used in the processing plant is well proven in many other operations locally and globally.  The processing testwork is representative of the different material types throughout the mining area  No deleterious material has been identified  The metallurgical recovery for gold is assigned by material type based on testwork: Oxide –94%

45 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
	 For minerals that are defined by a specification, has the ore reserve estimation been based on the appropriate mineralogy to meet the specifications?	Transitional – 93% Granite – 91% Mafic – 91%
Environment	 The status of studies of potential environmental impacts of the mining and processing operation. Details of waste rock characterisation and the consideration of potential sites, status of design options considered and, where applicable, the status of approvals for process residue storage and waste dumps should be reported.	 No environmental issues are known to exist which will prevent open pit mining taking place. Perseus has sufficient space available for waste dumps to store the expected quantities of mine waste rock associated with the Bélé open pit Ore Reserve. Based on the mineralogy of the deposit no potentially acid forming material has been identified.  Bélé does not require a standalone tailings storage facility as the ore material will be processed at Sissingué. An appropriate allowance has been made for the incremental tailings capacity required at Sissingue.
Infrastructure	 The existence of appropriate infrastructure: availability of land for plant development, power, water, transportation (particularly for bulk commodities), labour, accommodation; or the ease with which the infrastructure can be provided, or accessed.	 Activities in Bélé will be mining only and will be supported from the Sissingue operation.  Mining contractor will source the majority of labour from nearby villages.  There is an existing road from Bélé to Sissingue which in some sections will require upgrade.  Small creek diversions will be put in place at both Bélé East and Bélé West.
Costs	 The derivation of, or assumptions made, regarding projected capital costs in the study.  The methodology used to estimate operating costs.  Allowances made for the content of deleterious elements.  The derivation of assumptions made of metal or commodity price(s), for the principal minerals and co- products.  The source of exchange rates used in the study.  Derivation of transportation charges.  The basis for forecasting or source of treatment and refining charges, penalties for failure to meet specification, etc.  The allowances made for royalties payable, both Government and private.	 The mining costs are based on a schedule of rates provided to Perseus by tender quotes for mining contractors at the Sissingué. The same mining fleet can be used at Bélé and so the rates from Sissingué are considered appropriate  Allowance has been made for haulage of ore from Bélé to Sissingué (60km), based on rates currently being incurred for haulage of construction materials over significant distances for construction at Sissingué.  No deleterious materials have been identified.  Gold is the only metal considered in the Ore Reserves.  All costs are in US$.  A Bullion and Refining cost of US$3.6/oz was applied.  A royalty of 4.5% of the metal price was applied.
Revenue factors	 The derivation of, or assumptions made regarding revenue factors including head grade, metal or commodity price(s) exchange rates, transportation and treatment charges, penalties, net smelter returns, etc.  The derivation of assumptions made of metal or commodity price(s), for the principal metals, minerals and co- products.	 A Gold price of US$1,200/oz was used for calculating the revenue and validated using published metal price forecasts.  A Bullion and Refining cost of US$3.6oz was applied  A royalty of 4.5% of the metal price was applied
Market assessment	 The demand, supply and stock situation for the particular commodity, consumption trends and factors likely to affect supply and demand into the future.  A customer and competitor analysis along with the identification of likely market windows for the product.	 The demand for gold is considered in the gold price used.  It was considered that gold will be marketable for beyond the processing life.  The processing forecast and mine life are based on life of mine plans.  The commodity is not an industrial metal

46 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
	 Price and volume forecasts and the basis for these forecasts.  For industrial minerals the customer specification, testing and acceptance requirements prior to a supply contract.
Economic	 The inputs to the economic analysis to produce the net present value (NPV) in the study, the source and confidence of these economic inputs including estimated inflation, discount rate, etc.  NPV ranges and sensitivity to variations in the significant assumptions and inputs.	 A Life of Mine plan and economic model has been completed using the Ore Reserves published in this Statement. The inputs used are as per those stated in the relevant sections of this Statement. The assessment used a discount rate of 10%.  The NPV and sensitivity analysis outcomes indicated that the Project is economically viable. A schedule and economic model has been completed by Perseus on a pre- tax basis using the Ore Reserves published in this Statement. The inputs used are as per those stated in the relevant sections of this Statement. The assessment used a discount rate of 10% which is considered appropriate by RPM
Social	 The status of agreements with key stakeholders and matters leading to social licence to operate.	 Perseus will establish relevant agreements with local stakeholders using the model implemented successfully at Sissingué.  The plan for the operation of the Bélé open pits includes the use of skilled expatriate workers and locally sourced skilledworkers.
Other	 To the extent relevant, the impact of the following on the project and/or on the estimation and classification of the Ore Reserves:  Any identified material naturally occurring risks.  The status of material legal agreements and marketing arrangements.  The status of governmental agreements and approvals critical to the viability of the project, such as mineral tenement status, and government and statutory approvals. There must be reasonable grounds to expect that all necessary Government approvals will be received within the timeframes anticipated in the Pre-Feasibility or Feasibility study. Highlight and discuss the materiality of any unresolved matter that is dependent on a third party on which extraction of the reserve is contingent.	 The estimate of Ore Reserves for the Bélé open pits are not materially affected by any known environmental, permitting, legal, title, taxation, socio- economic, marketing, political or other relevant factors that would prevent the classification of Ore Reserves.
Classification	 The basis for the classification of the Ore Reserves into varying confidence categories.  Whether the result appropriately reflects the Competent Person’s view of the deposit.  The proportion of Probable Ore Reserves that have been derived from Measured Mineral Resources (if any).	 Ore Reserves have been classified based on the underlying Mineral Resources classifications and the level of detail in the mine planning. The Mineral Resources were classified as Measured, Indicated and Inferred. The Ore Reserves, based only on the Measured and Indicated Resources, have been classified as Proven and Probable Ore Reserves, respectively.  The Ore Reserve is classified as Proved and Probable in accordance with the JORC Code, corresponding to the Mineral Resource classifications of Measured and Indicated and taking into account other factors where relevant. The deposit’s geological model is well constrained. The Ore Reserve classification is considered appropriate given the nature of the deposit, the moderate grade variability, drilling density, structuralcomplexity andmining

47 | P a g e

==> picture [120 x 37] intentionally omitted <==

Criteria	JORC Code explanation	Commentary
		history. Therefore it was deemed appropriate to use Measured Mineral Resources as a basis for Proven Reserves and Indicated Mineral Resources as a basis for Probable Reserves.  No Inferred Mineral Resources were included in the Ore Reserve estimate.
Audits or reviews	 The results of any audits or reviews of Ore Reserve estimates.	 Perseus has completed an internal review of the Ore Reserve estimate.  The JORC Code provides guidelines which set out minimum standards, recommendations and guidelines for the Public Reporting of exploration results, Mineral Resources and Ore Reserves. Within the JORC Code is a "Checklist of Assessment and Reporting Criteria" (Table 1 - JORC Code). This checklist has been used as a systematic method to undertake a review of the underlying Study used to report in accordance with the JORC Code.  A high level LOM Plan was prepared based on the mineable ore contained with the pit designs. LOM Plan was reviewed for reasonableness and accuracy and confirmed that it was suitable for estimation of Ore Reserves. An economic model was prepared by Perseus that confirmed the operation to be economically viable.
Discussion of relative accuracy/ confidence	 Where appropriate a statement of the relative accuracy and confidence level in the Ore Reserve estimate using an approach or procedure deemed appropriate by the Competent Person. For example, the application of statistical or geostatistical procedures to quantify the relative accuracy of the reserve within stated confidence limits, or, if such an approach is not deemed appropriate, a qualitative discussion of the factors which could affect the relative accuracy and confidence of the estimate.  The statement should specify whether it relates to global or local estimates, and, if local, state the relevant tonnages, which should be relevant to technical and economic evaluation. Documentation should include assumptions made and the procedures used.  Accuracy and confidence discussions should extend to specific discussions of any applied Modifying Factors that may have a material impact on Ore Reserve viability, or for which there are remaining areas of uncertainty at the current study stage.  It is recognised that this may not be possible or appropriate in all circumstances. These statements of relative accuracy and confidence of the estimate should be compared with productiondata,where available.	 The accuracy and confidence of the inputs are, as a minimum, of a pre-feasibility level (for the global open pit Ore Reserves).  The key factors that are likely to affect the accuracy and confidence in the Ore Reserves are: o Accuracy of the underlying Resource Block Models; o Changes in Gold prices and sales agreements; o Changes in metallurgical recovery; and o Mining loss and dilution.  The Ore Reserve has utilised all parameters provided by testwork, tenders on key costs and prefeasibility level engineering assessment.  The accuracy of the underlying Mineral Resources is defined by the Resource Classification that the Mineral Resources are assigned to. Only the highest categories of Resource Classification i.e. Measured and Indicated have been used as a basis for estimating Ore Reserves.

48 | P a g e