MINERAL RESOURCE ESTIMATE

UPDATE ON THE BORATE

PROJECT, , OCTOBER 2016

Prepared For Erin Ventures Inc.

Report Prepared by

SRK Consulting (UK) Limited UK6467

name to work with template script in local application 030613 SRK Consulting Piskanja MRE 2016

COPYRIGHT AND DISCLAIMER

Copyright (and any other applicable intellectual property rights) in this document and any accompanying data or models which are created by SRK Consulting (UK) Limited ("SRK") is reserved by SRK and is protected by international copyright and other laws. Copyright in any component parts of this document such as images is owned and reserved by the copyright owner so noted within this document. The use of this document is strictly subject to terms licensed by SRK to the named recipient or recipients of this document or persons to whom SRK has agreed that it may be transferred to (the “Recipients”). Unless otherwise agreed by SRK, this does not grant rights to any third party. This document shall only be distributed to any third party in full as provided by SRK and may not be reproduced or circulated in the public domain (in whole or in part) or in any edited, abridged or otherwise amended form unless expressly agreed by SRK. Any other copyright owner‟s work may not be separated from this document, used or reproduced for any other purpose other than with this document in full as licensed by SRK. In the event that this document is disclosed or distributed to any third party, no such third party shall be entitled to place reliance upon any information, warranties or representations which may be contained within this document and the Recipients of this document shall indemnify SRK against all and any claims, losses and costs which may be incurred by SRK relating to such third parties. This document is issued subject to the confidentiality provisions in SRK‟s Terms and Conditions, which are included in the Commercial Appendices and contain mutual confidentiality obligations. Accordingly, any references in the confidentiality provisions in SRK‟s Terms and Conditions to the “Client” should be read as “Recipients”. SRK respects the general confidentiality of its potential clients‟ confidential information whether formally agreed with them or not and SRK therefore expects the contents of this document to be treated as confidential by the Recipients. The Recipients may not release the technical and pricing information contained in this document or any other documents submitted by SRK to the Recipients, or otherwise make it or them available to any third party without the express written consent of SRK.

© SRK Consulting (UK) Limited 2016 version: Jan2016

SRK Legal Entity: SRK Consulting (UK) Limited SRK Address: 5th Floor Churchill House 17 Churchill Way Cardiff, CF10 2HH Wales, United Kingdom.

Date: November, 2016

Project Number: UK6467

SRK Project Director: Mike Armitage Chairman and Corporate Consultant SRK Project Manager: Mike Armitage Chairman and Corporate Consultant

Erin Ventures I9nc Client Legal Entity:

Client Address: Suite 203 645 Fort Street Victoria, BC Canada V8W 1G2

U6467 Piskanja MRE 2016_Final Report.docx November, 2016

SRK Consulting (UK) Limited 5th Floor Churchill House 17 Churchill Way City and County of Cardiff CF10 2HH, Wales United Kingdom E-mail: [email protected] URL: www.srk.co.uk Tel: + 44 (0) 2920 348 150 Fax: + 44 (0) 2920 348 199 EXECUTIVE SUMMARY MINERAL RESOURCE ESTIMATE UPDATE ON THE PISKANJA BORATE PROJECT, SERBIA, OCTOBER 2016

1 EXECUTIVE SUMMARY 1.1 Introduction

SRK Consulting (UK) Ltd (SRK) has been requested by Erin Ventures Inc. (Erin or the Company) to prepare an updated Mineral Resource Estimate (MRE) for the Piskanja Borate Project (“Piskanja” or “the Project”).

During 2015, the Company undertook a programme of exploration work aimed at tightening the drill hole spacing at Piskanja to an approximate 50m by 50m grid within the central area of the deposit. The aim of this was to improve confidence to the block grade and tonnage estimates and to hopefully both confirm the previous MRE also produced by SRK in November 2013 and move some of the mineralisation into a higher confidence classification.

The responsible person for the updated MRE is Dr Mike Armitage who is a Qualified Person in accordance with the CIM Definition Standards on Mineral Resources and Reserves (CIM Standards).

1.2 Project Description

The Project occurs within a 306 hectare exploration licence located in southern Serbia, some 160 km south of the Serbian capital Belgrade. Balkan Gold doo, which holds the Exploration Licence, is a wholly owned subsidiary of the Company.

1.3 Project Geology

Geologically, the Project is located within the Jarandol basin, a Neogene continental sedimentary basin located within the Vardar Zone (VZ) tectonic belt. The host rocks comprise mudstones, sandstones, carbonates, tuffaceous sediments and conglomerates. It is interpreted that hydrothermal and tectonic activity led to borate mobilisation and deposition as a series of stratiform lenses within the basinal carbonate sediments.

Mineralised horizons show a range of textures predominantly comprising different growths of the borate minerals colemanite, ulexite and howlite.

Registered Address: 21 Gold Tops, City and County of Newport, NP20 4PG, Group Offices: Africa Wales, United Kingdom. Asia SRK Consulting (UK) Limited Reg No 01575403 (England and Wales) Australia Page 3 of 88 Europe North America

South America SRK Consulting Piskanja MRE 2016 –Executive Summary

1.4 Exploration Drilling and Sampling

The updated MRE presented here is based on some 32,880 m of drilling for a total of 98 drillholes. The drilling has all been completed from the surface on a grid spacing of approximately 50–100 m, providing intersections at a similar spacing. Drillholes are typically vertical and intersection angles with the mineralisation typically range from perpendicular to 45°.

In comparison to the MRE reported in November 2013 (the 2013 MRE), the database includes an additional 12 drillholes for 3,458 m of diamond drilling. During the recent exploration programme, samples were sent for preparation to SGS Laboratories sample preparation facility in Bor, Serbia, and then dispatched to SGS Lakefield, Canada for analysis

for boron by Na2O2-fusion ICP-OES.

SRK is confident that the data provided by the Company is of sufficient quality, and has been subjected to a sufficiently high level of verification to support the MRE as presented here.

1.5 Mineral Resource Estimate

In summary, in producing the updated MRE, SRK has:-

 Modelled the borate horizons in 3D;

 Composited the sample data to 1m intervals and undertaken a statistical analysis of the assay data in each mineralised domain;

 Evaluated the composited assay data for the presence of high-grade outliers from histograms;

 Undertaken geostatistical analyses to determine appropriate interpolation algorithms;

 Created a block model with block dimensions of 10x10x2 m;

 Undertaken a Quantitative Kriging Neighbourhood Analysis (QKNA) to test the sensitivity of the interpolation parameters;

 Interpolated borate grades into the block model;

 Visually and statistically validated the interpolated block grades relative to the original sample results; and

 Reported an MRE according to CIM Standards.

Upon consideration of data quality, drill hole spacing and the interpreted continuity of grades controlled by the deposit, SRK has variously classified portions of the deposit in the Indicated and Inferred Mineral Resource categories.

SRK has applied basic economic considerations to restrict the Mineral Resource to material that it considers has reasonable prospects for economic extraction by underground mining methods. This is primarily based on the results of SRK‟s previous 2014 Preliminary Economic Assessment (PEA) though SRK has reviewed this in the light of the updated geological model produced for this estimate.

SRK‟s updated MRE for the Project is presented in Table ES 1 below.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page iv of iii SRK Consulting Piskanja MRE 2016 –Executive Summary

Table ES 1: SRK Mineral Resource Statement as at 19 July 2016 for the Piskanja Deposit prepared in accordance with CIM Standards

Category Cut-off Tonnes Mt B2O3 Grade % B2O3 Mt Indicated 7.8 31.0 2.4 12% B2O3 Inferred 3.4 28.6 1.0

1.6 Comparison to Previous Estimates

In comparison to the previous SRK 2013 MRE for the Project, which was also reported at a

cut-off grade of 12% B2O3 but above a minimum mining thickness of 1.0 m, this updated MRE (which is reported above a minimum mining thickness of 1.2 m) has more borate in the Indicated category (2.4Mt compared to 1.7 Mt) and less borate in the Inferred category (1.0Mt compared to 1.8 Mt). These changes are primarily due to some previously reported Inferred material being upgraded in to the Indicated category though there is also a slight reduction in

B2O3 content for the Project as a whole from 3.5 Mt to 3.4 Mt (-4%) mainly as a result of a 6% reduction in tonnage. This change is primarily due to the infill drilling which has shown some of the thicker zones to be less continuous than previously assumed but also enabled the higher grade zones to be better delineated, which has reduced the tonnage slightly but at the same time slightly improved the grade (namely a global 2% increase in grade compared to the previous estimate).

1.7 Conclusions

The infill drilling and database verification work completed since the last MRE was produced has added further confidence to the geological model, enabled the borate distribution to be interpolated with more confidence grade distributions and allowed the reporting of an updated MRE which is similar in quantum, but is now more robust, than that previously produced.

There are, however, still areas of lower geological confidence which require more drilling and which may be subject to further revision in the future and there is also some associated work which could be done alongside or prior to more drilling to improve confidence in the estimate generally. This work is set out below.

1.8 Recommendations

Given the overall relatively minor change in global mean grade and tonnage estimate for the Project following the recent drilling, SRK recommends that further work should now focus on improving confidence in the geological and grade continuity through infill drilling at a mining scale, namely 10 or 20 m, in selected key areas of the deposit.

Improvements to data quality could also be achieved by:

 Completing additional density determinations to increase the number of results available for analysis and the confidence in the density model for the Project. SRK strongly recommends that any additional density sampling should record the weight of the sample following oven-drying prior to immersion in water, given the potential for moisture content to affect the density readings in the current database;

 Re-assaying by fusion of the following low-grade samples within the mineralised zones, which appear anomalously low compared with the adjacent drillhole intercepts: sample 2506 from EVP2015-137 and samples 2355 to 2356 from EVP2011-108;

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page v of iii SRK Consulting Piskanja MRE 2016 –Executive Summary

 Submitting the same 20 samples sent to SGS Lakefield (titration) and SGS Ankara (fusion) to another certified laboratory (for assay by fusion) to further investigate the bias (+8%) noted in the verification duplicates at SGS Ankara;

 Completing an additional phase of Round Robin analysis on the standard materials to further reduce uncertainty in the expected values and standard deviations and obtain external industry certification for the analytical results;

 Rectifying the sample interval overlaps and sample sequence, transcription and B% to

B2O3% conversion errors in the Company‟s assay database to reduce the need to re- validate the assays during future MRE updates. Furthermore, SRK recommends that the Company compiles a single clean interval table file in which to store the assays for boron, with separate columns reflecting different assay methodologies and a „final‟ assay field to ensure an easily-auditable database. Acquiring a commercial database storage system would assist with ensuring that the database is fully validated;

 Completing further work to fully understand the extent of sub-horizontal zones of faulting affecting the upper borate horizons (KZONE3 and KZONE4) and, significantly, the geotechnical implications of these for mining of these horizons; and

 Completing a drillhole re-logging exercise to highlight any borate intervals affected by or directly flanked by areas of sediment slumping, given the significance of this in terms of the short-scale geological continuity for mining, and also to incorporate the defining marker units in the stratigraphy (such as conglomerate horizons), given their use in adding confidence to the overall form and continuity of the stratiform mineralisation.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page vi of iii SRK Consulting Piskanja MRE 2016 – Table of Contents Main Report

Table of Contents 1 INTRODUCTION ...... 1 1.1 Background ...... 1 1.2 Requirement, Structure and Compliance ...... 1 1.3 Details of Personal Inspections ...... 1 1.4 Limitations, Reliance on SRK, Declaration, Consent, Copyright ...... 2 2 RELIANCE ON OTHER EXPERTS ...... 2 3 PROPERTY DESCRIPTION AND LICENCE LOCATION ...... 3 3.1 Project Location ...... 3 3.2 Mineral Licence Tenure ...... 4 3.3 Mining Rights in Serbia ...... 6 3.4 Surface Rights ...... 7 3.5 Permits and Authorisation ...... 7 3.6 Environmental Considerations ...... 8 3.7 Agreements and Royalties ...... 8 4 ACCESSIBILITY, CLIMATE, LOCAL RESOURCES, INFRASTRCUTURE AND PYSIOGRAPHY ...... 9 4.1 Accessibility ...... 9 4.2 Local Resources and Infrastructure ...... 9 4.3 Climate ...... 9 4.4 Physiography ...... 10 5 HISTORY ...... 11 5.1 History of Exploration and Mining ...... 11 5.2 Historical Mineral Resource Estimates ...... 11 5.3 Historical Production ...... 12 6 GEOLOGICAL SETTING AND MINERALISATION ...... 12 6.1 Background ...... 12 6.2 Regional Setting ...... 12 6.3 Stratigraphy...... 13 6.4 Structural Geology ...... 15 6.4.1 Tectonic Setting ...... 15 6.4.2 Regional Structures ...... 16 6.4.3 Project Scale Faulting ...... 16 6.4.3 Slumping ...... 17 6.5 Mineralisation ...... 18 6.5.1 Introduction ...... 18 6.5.2 Mineralisation Textures ...... 19 6.5.3 Mineralisation Geometry ...... 21 7 DEPOSIT TYPE ...... 22

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page i of v SRK Consulting Piskanja MRE 2016 – Table of Contents Main Report

8 ERIN EXPLORATION ...... 23 9 DRILLING ...... 23 9.1 Introduction ...... 23 9.2 Mines (1987-1992) ...... 23 9.3 Ras Borati (1997) ...... 24 9.4 Rio Tinto (2006-2007) ...... 24 9.5 Erin (2010- 2016) ...... 24 9.5.1 Overview ...... 24 9.5.2 Collar Surveys ...... 25 9.5.3 Downhole Surveys...... 25 9.5.4 Hole Orientation ...... 26 9.5.5 Diamond Drilling Procedure ...... 26 9.5.6 Core Recovery ...... 26 9.5.7 Core Storage ...... 27 9.6 SRK Comments ...... 28 10 SAMPLE PREPARATION, ANALYSIS AND SECURITY ...... 28 10.1 Introduction ...... 28 10.2 Diamond Drilling Sample Preparation and Chain of Custody ...... 28 10.3 Sample Preparation and Analysis ...... 28 10.4 Specific Gravity Data ...... 29 11 DATA VERIFICATION ...... 30 11.1 Introduction ...... 30 11.2 Assay QA/QC ...... 31 11.2.1 1987 - 2007 ...... 31 11.2.2 2011 - 2012 ...... 31 11.2.3 2015 ...... 33 11.3 Verifications by SRK ...... 35 11.3.1 General verification...... 35 11.3.2 Sample Database ...... 36 11.3.3 Assay Technique ...... 37 11.3.4 Non-sampled Intervals ...... 38 11.4 SRK Comments ...... 38 12 MINERAL RESOURCE ESTIMATES ...... 39 12.1 Introduction ...... 39 12.2 Statistical Analysis – Raw Data ...... 39 12.3 3D Modelling ...... 40 12.3.1 Mineralisation Domains ...... 40 12.3.2 Statistical Analysis ...... 41 12.3.3 Mineralisation Model Coding ...... 42

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page ii of v SRK Consulting Piskanja MRE 2016 – Table of Contents Main Report

12.4 Compositing ...... 44 12.5 Evaluation of Outliers ...... 44 12.6 Geostatistical Analysis ...... 45 12.7 Block Model and Grade Interpolation ...... 45 12.8 Final Interpolation Parameters ...... 46 12.9 Model Validation and Sensitivity ...... 47 12.9.1 Sensitivity Analysis ...... 47 12.9.2 Block Model Validation ...... 47 12.10 Mineral Resource Classification ...... 50 12.11 Mineral Resource Statement ...... 51 12.12 Grade Sensitivity Analysis ...... 52 12.13 Comparison to Previous Mineral Resource Estimates ...... 52 13 INTERPRETATIONS AND CONCLUSIONS ...... 53 14 RECOMMENDATIONS ...... 53 15 REFERENCES ...... 54

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page iii of v SRK Consulting Piskanja MRE 2016 – Table of Contents Main Report

List of Tables Table 3-1: The history of the validity of the Tenement covering the Piskanja Project area...... 4 Table 3-2: Licence boundary coordinates for the Piskanja Project, licence #1934 ...... 4 Table 3-3: Royalties due on various extracted minerals (from Law on Mining and Geological Researches, 2012) ...... 8 Table 9-1: Summary of Piskanja Drilling as at 31 May 2016 ...... 25 Table 10-1: Summary of density statistics ...... 29 Table 11-1: Summary of Standard Material for boron submitted by the Company in sample submissions ...... 32 Table 11-2: List of Drillholes Excluded from the 2016 MRE ...... 36 Table 12-1: Summary of Mineralisation Zones at the Piskanja Project ...... 43 Table 12-2: Composite Statistics for Borate ...... 44 Table 12-3: Modelled semi-variogram parameters for Domain (GROUP 100)* ...... 45 Table 12-4: Block Model Dimensions ...... 46 Table 12-5: Final Interpolation Parameters ...... 46 Table 12-6: Summary Block Statistics for Ordinary Kriging and Inverse Distance Weighting Interpolation Methods ...... 50 Table 12-7: SRK Mineral Resource Statement as at 19 July 2016 for the Piskanja Deposit prepared in accordance with the CIM Standards ...... 52 Table 12-8: Gradations for Open Pit Material at Piskanja at various Borate % Cut-off Grades ..... 52

List of Figures Figure 3-1: Location of the Piskanja licence area ...... 3 Figure 3-2: Geographical map of the Piskanja Exploration Licence #1934 (red line)...... 5 Figure 4-1: Examples of the terrain and agricultural land use typical of the licence area. Before drilling hole EVP2011-103 (left) and after drilling and remediation (right)...... 10 Figure 6-1: Regional E-W cross-section approximately coinciding with the location of the Piskanja Deposit (modified from Matenco & Radivojević 2012) ...... 13 Figure 6-2: Cross-section of the stratigraphy and structure of the Piskanja area (Erin 2013) ...... 14 Figure 6-3: 1:5,000 Geological Map of the Piskanja Project, (Erin 2013) ...... 15 Figure 6-4: Colour-shaded topographic map of the Piskanja area, showing SRK‟s interpretation of potential faults that may affect the Piskanja project (drillholes are red circles). Faults away from the deposit are omitted ...... 17 Figure 6-5: Chaotic soft-sediment deformation within a broadly concordant layer of claystones and sandstones, Bella Sten magnesite pit ...... 18 Figure 6-6: Examples of pre-lithification structures, both extensional and contractional in nature, affecting sandstone and mudstones of the Piskanja Project ...... 18 Figure 6-7: Mineralisation textures: (a) Remnant carbonate partings in massive borate; (b) Layer- parallel vein showing vertical opening direction; (c) Variably brecciated interval; (d) Primary vug; (e) Secondary dissolution vug ...... 20 Figure 6-8: Massive borate mineralisation in hole EVP2012-111 from 310.30 m to 313.20 m, situated at the contact between shale and dolomite units ...... 21 Figure 6-9: Plan view of true thickness for borate horizon KZONE3 ...... 21 Figure 9-1: Location of new collars (pink) completed by the Company during the 2015 exploration program ...... 25 Figure 9-2: Example cross section through the Piskanja deposit ...... 26 Figure 9-3: Core Recovery within mineralised horizons at the Piskanja Project ...... 27 Figure 9-4: Core logging and storage facility (2012)...... 27 Figure 10-1: Sample bags prepared for transport to the laboratory ...... 28 Figure 11-1: Composite sample grade histogram distributions for borate, showing data assayed with QAQC support (left) and without QAQC (right) ...... 30 Figure 11-2: QAQC Standard Summary Charts from submission of Piskanja Samples (2011/2012) showing analysis by titration (left) and ARD (right) ...... 33 Figure 11-3: QAQC Standard Summary Charts from submission of Piskanja Samples (2015) showing analysis for boron using Na2O2-fusion ...... 35 Figure 11-4: Scatter plot for B2O3% samples (170) analysed by titration and ARD ...... 37 Figure 11-5: Scatter plot for B2O3% samples (82) analysed by titration and fusion ...... 38 Figure 12-1: Incremental Histogram of Length Weighted Project Borate Assays ...... 40

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page iv of v SRK Consulting Piskanja MRE 2016 – Table of Contents Main Report

Figure 12-2: Log histogram plot for borate for domain KZONE3 ...... 41 Figure 12-3: Assessment of 3D borate grade distribution within the KZONE3 domain ...... 42 Figure 12-4: 2D assessment of borate grade distribution downhole, looking west ...... 42 Figure 12-5: Piskanja Mineralisation Model, looking north ...... 43 Figure 12-6: Piskanja Mineralisation Model, looking down ...... 43 Figure 12-7: Raw and Log Histograms for borate for the KZONE3 domain ...... 44 Figure 12-8: Piskanja Block Model Borate Grade Distribution (3D view, looking north) ...... 48 Figure 12-9: Piskanja Block Model Borate Grade Distribution (cross-section) ...... 48 Figure 12-10: Validation Plot (Easting) showing Block Model Estimates versus Sample Mean (40m Intervals) for domain KZONE1 for borate ...... 49 Figure 12-11: SRK Mineral Resource Classification...... 51

List of Technical Appendices A QAQC ANALYSIS ...... A-1 B MINERAL RESOURCE ESTIMATE ...... B-1

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page v of v SRK Consulting (UK) Limited 5th Floor Churchill House 17 Churchill Way City and County of Cardiff CF10 2HH, Wales United Kingdom E-mail: [email protected] URL: www.srk.co.uk Tel: + 44 (0) 2920 348 150 Fax: + 44 (0) 2920 348 199 MINERAL RESOURCE ESTIMATE UPDATE ON THE PISKANJA BORATE PROJECT, SERBIA, OCTOBER 2016

1 INTRODUCTION 1.1 Background

SRK Consulting (UK) Ltd (SRK) has been commissioned by Erin Ventures Inc. (Erin, hereinafter also referred to as the Company) to prepare an updated Mineral Resource Estimate (MRE) for the Piskanja Borate Project (“Piskanja” or “the Project”). Balkan Gold doo, which holds the Exploration Licence, is a wholly owned subsidiary of the Company. The Company is currently listed on the Toronto Stock Exchange Venture Exchange (TSX-V) using the code EV.

The Project comprises a 306 hectare concession package located in southern Serbia, some 160 km south of the Serbian capital Belgrade.

SRK first produced a MRE for the Project in November 2013 and provides this July 2016 update based on further infill drilling and database verification. Specifically this updated MRE is based on some 32,880 m of drilling for a total of 98 drillholes.

The latest phase of exploration work completed by the Company has been focused towards increasing the sample coverage to 50x50m within the central area of the deposit, with an aim of adding further confidence to the Project‟s block grade and tonnage estimates.

1.2 Requirement, Structure and Compliance

The reporting standards adopted for the reporting of the MRE are the Canadian Institute of Mining, Metallurgy and Petroleum (“CIM”) Definition Standards for Mineral Resources and Mineral Reserves (adopted May 2014) (CIM Standards). This is an internationally recognised reporting code as defined by the Combined Reserves International Reporting Standards Committee (CRIRSCO).

The Qualified Person (“QP”) responsible for this report is Dr Mike Armitage who assumes responsibility for the geological model, resource estimation procedures and the report as a whole.

1.3 Details of Personal Inspections

This report is based on information collected by SRK during site visits performed during the course of 2012, 2014 and 2016 and on additional information provided by the Company throughout the course of SRK‟s investigations. Other information was obtained from the public domain.

Registered Address: 21 Gold Tops, City and County of Newport, NP20 4PG, Group Offices: Africa Wales, United Kingdom. Asia SRK Consulting (UK) Limited Reg No 01575403 (England and Wales) Australia Page 1 of 88 Europe North America South America SRK Consulting Piskanja MRE 2016 – Main Report

1.4 Limitations, Reliance on SRK, Declaration, Consent, Copyright

SRK‟s opinion contained herein and effective 19 July 2016, is based on information collected by SRK throughout the course of SRK‟s investigations, which in turn reflect various technical and economic conditions at the time of writing. Given the nature of the mining business, these conditions can change significantly over relatively short periods of time and consequently the MRE presented here may similarly change with time.

This report may include technical information that requires subsequent calculations to derive sub-totals, totals and weighted averages. Such calculations inherently involve a degree of rounding and consequently introduce a margin of error. Where these occur, SRK does not consider them to be material.

SRK is not an insider, associate or an affiliate of the Company, and neither SRK nor any affiliate has acted as advisor to the Company, its subsidiaries or its affiliates in connection with this project. The results of the technical review by SRK are not dependent on any prior agreements concerning the conclusions to be reached, nor are there any undisclosed understandings concerning any future business dealings.

Except as specifically required by law, SRK does not assume any responsibility and will not accept any liability to any other person for any loss suffered by any such other person as a result of, arising out of, or in connection with this Technical Report or statements contained herein, required by and given solely for the purpose of complying with the mandate as outlined in this Technical Report. SRK has no reason to believe that any material facts have been withheld by the Company.

This report is intended to be read as a whole, and sections should not be read or relied upon out of context. The technical report contains expressions of the professional opinion of the Qualified Person based upon information available at the time of preparation.

2 RELIANCE ON OTHER EXPERTS

The majority of the information reviewed in preparing this report has been provided by the Company. SRK has referenced information and data sourced from reports and documents where applicable.

Notably, SRK has relied on the following technical reports and information:

 2006, Geosystem srl, Magnetotelluric Survey, Jarandol Basin, Serbia;

 2012, University of Belgrade, Faculty of Mining and Geology, Testing of samples from the Piskanja borate deposit (translation from Serbian);

 2012, University of Belgrade, Faculty of Mining and Geology, Petrological characteristics of holes 104, 105, 106, 107, IBM-4 and IBM-6 – Piskanja (in Serbian);

 2012, SGS Minerals Services, Report on magnetic and HTE testing of borate samples from Serbia;

 2013, University of Belgrade, Faculty of Mining and Geology, Study of engineering properties rock masses and terrains of the Piskanja borate deposit (translation of concluding remarks from Serbian), and;

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 2 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

 2014, SRK Consulting UK Ltd., NI43-101 Technical Report and Preliminary Economic Assessment for the Piskanja Borate Project, Serbia

 2016, SRK Consulting UK Ltd., An Evaluation of the Structural Geology of the Piskanja Borate Project, Serbia

3 PROPERTY DESCRIPTION AND LICENCE LOCATION 3.1 Project Location

The Project covers an area of 305.7 hectares. The approximate centre of the project area is 43º 22‟ 43”N and 20º 38‟ 50”E in standard degrees, minutes, seconds format. The Project is located in southern Serbia, some 160 km south of the Serbian capital Belgrade. Nearby towns include: , 40 km to the north; , 28 km to the south, and; Raška, 11 km to the south, (Figure 3-1).

Figure 3-1: Location of the Piskanja licence area

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 3 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

3.2 Mineral Licence Tenure

The Ministry of Mining and Environmental and Spatial Planning of the Republic of Serbia (the Ministry) first granted Balkan Gold doo Exploration Licence #1934 on 23 August 2010 under the 1995 Law on Mining published in Official Gazette of RS, no. 44/95.

In July 2012, Balkan Gold doo was granted a three year Exploration Licence under the Law on Mining and Geological Researches, published in Official Gazette no. 88/2011 which came into force in January 2012. Balkan Gold doo has most recently renewed the Exploration Licence #1934 on 11 July 2016, which is valid until 11 July 2018. The licensed area remains the same as the previous title (3.075 km2) and is defined by the coordinates in Table 3-2 and Figure 3-2. The licence covers the Piskanja mineral deposit in its entirety. There are no other known mineral deposits within the licence area.

On 10 December 2012, Balkan Gold doo was granted Exploration Licence #2065 which covers an area of 34.90 km2 and is situated adjacent to the Exploration Licence #1934. This licence was most recently renewed on 28 April 2016 and it is valid until 28 April 2018. The Company continues to explore Exploration Licence #2065 for boron mineralisation and associated elements (Li, Na, Sr and K).

This Technical Report is limited to work completed on Exploration Licence #1934.

Table 3-1: The history of the validity of the Tenement covering the Piskanja Project area. Licence Date Valid Tenement Name Date of Expiry Licence Area Number. from 08/23/2012 (extended from 1934 Piskanja 08/12/2010 306 ha 13/09/2011) 1934 Piskanja 05/11/2012 05/11/2015 306 ha 1934 Piskanja 11/07/2016 11/07/2018 306 ha

Table 3-2: Licence boundary coordinates for the Piskanja Project, licence #1934 Serbian Gauss Kruger - Zone 7 UTM_WGS 84_34N Point East_SGK North_SGK UTM_East_MI UTM_North_MI 1 7,471,000 4,803,000 470,575.04 4,802,071.89 2 7,471,000 4,804,750 470,575.04 4,803,821.36 3 7,472,750 4,804,750 472,324.51 4,803,821.36 4 7,472,750 4,803,000 472,324.51 4,802,071.89

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 4 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Figure 3-2: Geographical map of the Piskanja Exploration Licence #1934 (red line).

Balkan Gold doo‟s main responsibilities as licence holder are described in the “Decision of the Ministry of Natural Resources, Mining and Spatial Planning” dated 05 November 2012. It is understood by SRK that this decree states that Balkan Gold doo is committed to performing exploration activities in accordance with the Exploration Programme submitted to the Ministry at the time of licence application.

The Company contracted Ibarski Rudnici Coal Company, which is a subsidiary of the State- owned JP PEU Resavica, to design its most recent (2012-2015) Exploration Programme. In accordance with the 2012 Law on Mining and Geological Researches of Serbia, this Exploration Programme was approved by the Institute for Nature Conservation of Serbia and the Institute for Cultural Heritage and Preservation, Kraljevo, prior to it being submitted to the Ministry. Previously, the Company‟s 2010 Exploration Programme was designed and submitted by private exploration consultancy, Jantar Group, Belgrade.

The Company completed the following exploration activities during 2012-2015 to satisfy the requirements set by Law on Mining and Geological Researches of Serbia:

 The drilling of 7 holes to validate the pre-1997 exploration;

 Hydrogeological and hydrological studies (on-going);

 Mineralogical and petrological studies including SEM-EDS and XRD;

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 5 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

 Geotechnical studies of the core;

 Preliminary metallurgical testing using SGS and SCL laboratories (2 x25 kg);

 Analytical tests (ICP, titration and XRF) on all new core samples (completed on all holes drilled in 2011/2012);

 Maintenance of a GIS model and data base (on-going);

 Preparation of a Mineral Resource/Ore Reserve estimate report (Serbian report completed);

 The drilling of 11 further holes using a 50m x 50m grid of drill collars;

 Topographic surveying at 1:1000 scale over the exploitation area (150-200 ha) (completed during 2016); and

 The preparation of annual reports and a final report on geological exploration for the validity of the licence period.

The Company plans to complete a further 3 drill holes at 50x50m coverage during the next phase of infill drilling.

SRK understands that it is required by Serbian Law on Mining and Geological Researches that exploration activities and annual reports submitted to the Ministry must be monitored by a third party company. The following organisations have been responsible for such monitoring of the Company‟s on-going Exploration Programme, although it should be noted that SRK has not verified the listed reports or third party companies involved:

 Monitoring of the technical programme in 2010 was conducted by Geoprofesional Ltd, based in Belgrade;

 Technical monitoring of the programme and the annual report in 2011 were completed and submitted by Silur doo, based in Kraljevo; and

 Technical monitoring of the programmes and the annual reports in 2012, 2013, 2014 and 2015 were completed and submitted by South Danube Metals (a wholly-owned subsidiary of Euromax Resources Ltd), based in Belgrade.

3.3 Mining Rights in Serbia

The laws relating to Mining Rights in Serbia are described in the document titled “Law on Mining and Geological Researches” which is published in “Official Gazette of RS” #101/2015.

This document states that an Exploration Licence may be granted for an initial period of three years, and then be extended twice more for a further two years on each renewal. The total allowed exploration period is 7 years. Under Serbian Law, the Exploration Company must submit annual reports of the work completed which evidence that not less than 75% of the planned work has been completed.

According to Serbian Law, it is necessary to undertake a feasibility study prior to applying for a Mining Licence. The Company plans to complete all necessary steps in order to apply for a Mining Licence.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 6 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Article 70 of the Serbian Law on Mining and Geological Researches defines the items that must be addressed and attached to a Mining Licence application as:

 Proof of payment the republic, i.e. provincial administrative fee when the exploitation is carried out on the territory of the Autonomous Province;

 A topographic map in a scale 1:25000 or at corresponding scale with drawn‐in boundaries of the exploitation field and contours of determined reserves of mineral resources, public traffic roads and other facilities located in that field and clear marked cadastral plots in a written and digital form;

 A certificate on resources and reserves of mineral resources issued on the basis of performed explorations in accordance with applicable regulations on classification the resources and reserves;

 A certificate of registration and a copy of the appropriate act document indicating the activity codes for which the applicant is registered, the registration number of the company and the corresponding license;

 A feasibility study on the mining of the deposit; and

 A local government act confirming the compliance of exploitation with the appropriate spatial or urban plans and a possible need of development the planning document of lower rank.

3.4 Surface Rights

The Surface Rights over the Piskanja mineral deposit are held by private individuals and by local/state governments. Land access therefore is negotiated with the individual landowners, for which they are reimbursed according to a payment scheme organised by Balkan Gold doo.

The Company has informed SRK that some planned drill hole collar locations have on occasion had to be moved due to private land owners refusing access and while this has not affected the exploration programme to date, the progression to a tighter drill spacing will likely result in further such issues with landowners.

SRK understands that the Company intends, through Balkan Gold doo, to acquire the surface rights for a portion of municipal building land currently owned by the State-owned Ibarski Rudnici Coal Company for mining operations and construction.

3.5 Permits and Authorisation

Under the Serbian Law, a permit must be obtained from the relevant government department to ensure that known heritage sites are not impacted upon by exploration or mining activities. To satisfy this regulation an assessment was made by the Institute for Cultural Heritage Preservation, Kraljevo most recently on 5 November 2015 and a permit subsequently granted. The permit is valid until 5 November 2017.

Further permits may be required as the project develops and prior to commencing any mining operations.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 7 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

3.6 Environmental Considerations

Prior to Erin commencing exploration, site conditions were assessed by the Institute for Nature Conservation of Serbia and the Institute for Cultural Heritage Preservation, Kraljevo. An Environmental Permit was approved as a result. The Environmental Permit was most recently approved on 10 September 2015 and remains valid until 29 October 2017.

3.7 Agreements and Royalties

Articles 159 and 160 of the Serbian Law on Mining and Geological Exploration (Official Gazette RS No. 101/2015) state that companies undertaking mining activities shall pay a fee for the use of the mineral deposit. The Law states that “this revenue shall be the amount gained by the exploiting entity from used or natural mineral raw materials, determined on the basis of income gained from sale of non-refined mineral raw material, or income gained from the sale of technologically refined mineral raw material”. The fee will be split between the Republic of Serbia, the local government and the Ministry of Natural Resources, Mining and Spatial Planning.

As the law does not list the commodities classified under metallic and non-metallic minerals, it is not known whether the Ministry will impose a royalty on borates similar to that of other evaporite minerals such as gypsum (salt), or select to impose a levy specific to borate.

SRK recommends that the Company should contact the Ministry to confirm the royalty for borates for future studies.

Table 3-3 presents some of the Royalties due on certain commodities under Serbian Law.

Table 3-3: Royalties due on various extracted minerals (from Law on Mining and Geological Researches, 2012) Commodity Fee/Royalty All types of coal and oil shale 3% of income All metallic raw materials 5% of smelting plant net income Technogenic raw materials resulting from 1% of income exploitation and refining of mineral raw materials Non-metallic raw materials 5% of income All types of salts and salty solutions 1% of income

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 8 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

4 ACCESSIBILITY, CLIMATE, LOCAL RESOURCES, INFRASTRCUTURE AND PYSIOGRAPHY 4.1 Accessibility

The Project is located in the Jarandol Basin in the Raška region of south central Serbia approximately 160 km south of the Serbian capital Belgrade, and approximately 17 km north of the Kosovo border. The nearest settlement is the town of Baljevac na Ibru (literally “Baljevac on the Ibar River”) which is some 1.7 km northwest from the centre of the Exploration Licence #1934. Baljevac na Ibru (Baljevac) has a population of 1,482 (2011 census), (Figure 3-2). The regional capital, Raška, lies 10 km to the south of the Project area and has a population of 6,500 (2011 census).

Access to the Project is by paved road from Belgrade, a journey that takes approximately 4 hours and passes through the towns of Kragujevac and Kraljevo. Access around site is by vehicle/foot as the terrain is not steep and the land cleared for agriculture.

A standard gauge railway accommodating passenger and freight rolling stock passes through the western part of the licence area and runs from Belgrade, through the towns of Kraljevo and Raška to Poije in Kosovo.

4.2 Local Resources and Infrastructure

Erin has an office in the town of Baljevac, located at the Ibarski Rudnici coal mine, a small- scale operation exploited only for local coal supply. Facilities here also include core logging and sampling areas, and core and sample storage.

Water for exploration needs is sourced from streams that flow into the Ibar River and the water table is encountered in drillholes at shallow depths, for example at 23m below surface in hole EVP2012-100. It is reported that the stream water is used by the local people for drinking water. A 35 kV electricity lines run across the licence and supply power to Baljevac.

There is good mobile phone reception throughout the Jarandol Basin and the Project area.

4.3 Climate

The climate in the Project area is typical of Eastern Europe with four seasons of approximately equal length; spring, summer, autumn and winter. According to Foreca (an international company providing digital weather forecast data), the temperatures range from between 11 ºC and 28 ºC during the summer months of June to September, to between -3 ºC and 5 ºC during the middle winter months, December and January. Rainfall is highest in the months of May to September with the monthly average of between 49 and 62 mm. The months of January and February have the minimum amount of precipitation (about 30 mm), falling mostly as snow. Exploration activities can continue throughout the year with minimal inconvenience during the winter months.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 9 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

4.4 Physiography

The Jarandol Basin lies at an elevation of between 375 and 400 m above mean sea level (amsl), is elongated in an east-northeast - west-southwest direction and drains towards the north via the Ibar River which lies just outside the western boundary of the licence area. The terrain rises to approximately 750 m amsl to the west of the valley, outside of the Exploration Licence area, and to over 1,200 m immediately east of the Project area.

Minor tributaries to the Ibar River extend through the Exploration Licence area; the Kuricki to the North of the deposit and the Korlacki to the South. Between them is the Radic, an ephemeral water feature which is dry for most of the year.

The flood plains in the central part of the basin and the low angled valley sides are cultivated for crops and fruit, with the steeper terrain above 500 amsl generally covered by sparse deciduous woodland, (Figure 4-1).

As already commented, the Surface Rights over the Piskanja site are held by private individuals and or local/state governments. Land access has to therefore be negotiated with the individual landowners, for which they are reimbursed according to a payment scheme approved by the State. As stated in Section 3, the Company does not currently have any Surface Rights in the Project area and some private land owners have previously refused access for drilling. This has not affected the exploration programme to date although further infill drilling programmes will ideally require access to land where the owners have previously refused access. SRK understands that the Company intends to acquire the surface rights for a piece of industrial land currently owned by the State-owned Ibarski Rudnici Coal Company which it intends to use for mining operations and construction in the future.

Figure 4-1: Examples of the terrain and agricultural land use typical of the licence area. Before drilling hole EVP2011-103 (left) and after drilling and remediation (right).

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 10 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

5 HISTORY 5.1 History of Exploration and Mining

The first record of boron mineralisation in the Jarandol Basin was a hand-sized sample containing howlite found in a tributary of the Ibar river in 1967 during State-organised geological prospecting (Stojanovich, 1967). Following this, geological mapping at a scale of 1:10,000 was completed and the Pobrdje occurrence was identified some 2.6 km northwest of the Company‟s exploration licence.

The geochemical investigation of boron in the Jarandol Basin began in 1979 with the first identification of colemanite in a borehole occurring in 1987. Between 1987 and 1992, the Yugoslavian state-owned company Ibar Mines completed a number of soil and stream sediment sampling programmes, which were followed by an initial phase of drilling.

The Project was subsequently acquired by Ras Borati d.o.o (Ras Borati), a 50:50 joint venture company established between Erin Ventures Inc. and Elektroprevreda d.o.o., during which time (1997) a small number of Reverse Circulation (RC) holes were completed.

Following the resolution of international conflicts and a change of the governing party in 2006, Rio Tinto acquired the Project as part of its regional investigation of borate potential in Tertiary basins across the Balkan region. Further exploration work and drilling was completed 2006- 2007 including mineralogical investigations at the University of Belgrade and at Spectrum Petrographics Inc. of Vancouver, Canada, which identified the main boron-bearing minerals at the Project as colemanite and ulexite with minor howlite and probertite.

Rio Tinto also commissioned a magnetotelluric survey (Geosystem srl, 2006) to assess the conductivity variation within the Jarandol Basin sediments and to map the extent and thickness of the fine-grained sedimentary sequence. A low resistivity zone representing hydrous mineralisation was expected to be encountered. However, the results of the survey were considered inconclusive with respect to identifying conductivity variation related to borate mineralisation in the material close to surface (i.e. within 500m).

The Company acquired the exploration licence for the Project in August 2010 through its wholly owned subsidiary company, Balkan Gold doo.

5.2 Historical Mineral Resource Estimates

The Mining Institute of Belgrade prepared a mineral resource estimation report for Piskanja in 1992. This appears to have been an unofficial resource estimate as defined by Serbian law, created using cross sections between 200x300 m and 100x100 m spaced drill holes. The

mineral resource reported approximately 6.5 Mt of boric oxide (B2O3) in the C1+C2 categories as defined by Serbian mining regulations, though no records of grades were provided. SRK has not verified these estimates.

During the 2005 Public Tender of the Piskanja Project (Public Tender, 2005) the Ministry of Mining and Energy of the Government of the Republic of Serbia stated that „potential reserves of boron ore in Piskanja deposit were estimated to be 7,500,000 tonnes with an average

grade of 36.39% B2O3‟. As above, SRK has not verified this estimate.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 11 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

SRK previously produced a MRE for the Project with effective date of 29 November 2013.

This was reported above a 12% B2O3 cut-off and assuming a minimum mining height of 1 m,

and comprised an Indicated Mineral Resource of 5.6 Mt grading 30.8% B2O3 and an Inferred

Mineral Resource of 6.2 Mt grading 28.8% B2O3.

5.3 Historical Production

SRK is not aware of any significant production of borate from the Piskanja exploration permit.

6 GEOLOGICAL SETTING AND MINERALISATION 6.1 Background

The following sections describing the geological setting are sourced largely from descriptions provided in previous Technical Reports produced by SRK including SRK‟s 2014 PEA and SRK‟s 2016 Structural Geology report.

6.2 Regional Setting

Geologically, the Piskanja Deposit is located within the Jarandol basin, a Neogene continental sedimentary basin located within the Vardar Zone (VZ) tectonic belt. The VZ tectonic belt consists of ultrabasic blocks separated by fractured ophiolites that represent Early Mesozoic (Triassic-Jurassic) ophiolitic paleo-rifts. The western VZ ophiolitic unit represents a suture zone between the continental Adriatic Plate (Dinarides of the Western Serbia) and the European Plate (Carpatho-Balkanides and Macedonian Massif of Eastern Serbia). The structure of the Balkan region at this latitude, omitting the Neogene sediments, is shown as a cross-section in Figure 6-1, reproduced from Matenco & Radivojević (2012). The location of the Neogene sediments is broadly coincident with a series of tectonically interleaved nappes and obducted oceanic floor that were assembled in Cretaceous times, during the closure of the Neotethys ocean (Matenco & Radivojević 2012).

The Piskanja Deposit is located within sediments that developed within the Upper Jurassic ophiolite unit in the VZ tectonic belt during the Neogene (23-5 Ma). The development of the basin followed intense intermediate to acidic magmatism and granite plutonism in the Oligocene (34-23 Ma). These Neogene (mainly Miocene) basins in the VZ tectonic belt were continental in nature, their fills being a product of depositional processes associated with fluvial, lacustrine and swamp settings. The resulting sediments themselves comprise alternating units of mudstone, shale, sandstone and lignite caused by fluctuation in water depth and sediment input. Tuffaceous sediments related to on-going volcanism are also found within the basin and it is believed that related hydrothermal and tectonic activity led to borate mobilisation and deposition within the basinal sediments.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 12 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Figure 6-1: Regional E-W cross-section approximately coinciding with the location of the Piskanja Deposit (modified from Matenco & Radivojević 2012)

6.3 Stratigraphy

A semi-regional cross-section of the Piskanja area constructed by the Company is shown in Figure 6-2 which also summarises the local stratigraphy. The basement geology of the Jarandol Basin in this area comprises tectonised serpentinitic rocks relating to the Upper Jurassic ophiolites. Overlying these is a succession of andesitic to dacitic composition volcaniclastics, shown on Figure 6-2 to have a vertical thickness of 100-150 m in thickness.

The principal basin-fill recognised in the Piskanja Project area comprises conglomerates, carbonates, marls and tuffaceous sediments according to Obradovic et al. (1992). A coal deposit occurs close to the base of the succession, and is exploited just northeast of Baljevac at Odlagalište. Stratigraphically above the coal, but below the borate-bearing basin fill, a magnesite deposit (Mg-dolomite) has been exploited to the south of Baljevac at Bella Sten. It is unclear, however, precisely how far below the main borate-hosting horizons the magnesite deposits are.

Three main sedimentary packages are recognised by the Company in the locale of the Piskanja Project (Figure 6-3). These sediments comprise a total thickness of almost 560 m in places and are described below, from oldest to youngest:

 TcP1 (90-130 m): A conglomerate and sandstone unit, characterised by a dominance of coarse clastic sediments with a few thin interlayers of carbonate rocks. The thickness of individual layers of sedimentary breccias and conglomerates typically vary from 0.1 m to 10 m in general but can reach 25 m in the upper part of the unit.

 TcP2 (up to 330 m): A claystone and carbonate unit, characterised by thin (millimetre- scale) laminations of claystone, silty claystone, tuff, travertine, dolomite, dolomitic limestone with claystone and rarely sandstone, breccia and conglomerate. Metre-scale, bedding-concordant horizons of borate mineralisation are associated with the carbonate sediments.

 TcP3 (20-90 m): An upper claystone and sandstone interbedded unit. The sandstones are generally 1 to 2 m thick, with 2-10 m thick intervals of dolomitic carbonates. The claystone and sandstone are generally not laminated but possess a massive texture.

 Quaternary sediments (up to 25 m): Covering 65% of the licence area, these colluvial and alluvial sediments are characterised by the presence of rounded and semi-rounded pebbles and boulders mixed with fine and coarse sand.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 13 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

A

Figure 6-2: Cross-section of the stratigraphy and structure of the Piskanja area (Erin 2013)

*Line of cross section shown in Figure 6-3.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 14 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Figure 6-3: 1:5,000 Geological Map of the Piskanja Project, (Erin 2013)

6.4 Structural Geology

6.4.1 Tectonic Setting

Relatively little information on the tectonics of the Jarandol basin exists in scientific literature written in English. However, recent work by Matenco & Radivojević (2012) has summarised the main phases of tectonism for the Serbian part of the Pannonian Basin, including parts of the Jarandol Basin, which may reflect the local tectonism in the project area. According to these authors, although the precise timings and significance of each phase of tectonism varies across the basin, the basic framework of tectonic events is generally assumed to be as follows:

 Early Miocene (c. 20 Ma): Onset of extension.

 Middle Miocene times: Peak extensional tectonic activity along basinal normal faults.

 Late Miocene times: Post-rift thermal sag phase.

 Latest Miocene–Quaternary: Contractional event that overprinted the basin.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 15 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Apatite fission track dating of heating and cooling events in the area of the southern Ibar River Basin, including the Piskanja deposit, suggest the rocks may have attained a maximum temperature of 100-130°C between approximately 17 and 7 Ma due to uplift and transfer of heat from the Studenica- extensional core complexes (Andrić et al. 2015). This was followed by cooling of the basin, attributed to inversion. A thermal event at 7 Ma in the basin is tentatively attributed to a late phase of volcanism.

6.4.2 Regional Structures

Erin has completed a regional mapping campaign and has mapped fault zones throughout the wider area based on direct and indirect evidence of faulting. As can be seen from Figure 6-3, a map of the region mapped by Erin, the major structural trends comprise NW-SE and NE-SW trends, with sub-ordinate E-W and N-S trending faults. SRK has not been able to fully verify the evidence for this interpretation, but has seen good evidence from road-side outcrops along the Ibar valley of the presence of faults striking NW-SE, N-S and NE-SW.

Field observations made by SRK along the Ibar River valley, indicate that the NW-SE and N-S trending faults are likely relatively steep in nature. The Company‟s maps do not show resolvable fault displacements and it is likely that many structures represent faults with relatively small displacement (i.e. less than a few tens of metres).

A cross-section through the deposit in a NNE-SSW orientation has been tentatively interpreted by Erin to show synclinal folding of the Miocene basin (Figure 6-3). This may, however, simply reflect the presence of relatively localised sub-basins.

6.4.3 Project Scale Faulting

Within the immediate area of the Piskanja mineralisation, outcrop is limited and topography is generally rounded in nature and partially overlain by Quaternary sediments reaching up to ~30m in thickness. Therefore, the topographic expression of fault structures that are normally recessive in nature are covered in the area of borate mineralisation.

SRK concludes that there are no obvious major faults affecting the deposit. Notwithstanding this, based on the presence of recessive erosional features to the east and southeast of the deposit, SRK has produced a much-simplified and preliminary interpretation of potentially significant steeply dipping structures within the area of Piskanja (Figure 6-4). Indirect evidence of faulting supporting the orientation and position of with these structures include systematic changes in stratigraphic dip, the occurrence of a zone of thicker mineralisation, and the distribution of a conglomeratic-sandstone body within the basin. Due to data constraints and the relatively wide-spacing of drilling at present, these are all tentative interpretations and require further testing.

In addition (and more significantly), there are zones of brittle deformation within the core which suggest certain parts of the upper borate horizons (KZONE3 and KZONE4) are affected by sub-horizontal zones of faulting which anastomose in and out of the horizons and, in places, affect the entire borate horizon.

Further work would be needed to fully understand the extent of faulting at a mining scale and, most significantly, its geotechnical implications in terms of the mining of these horizons in particular.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 16 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Figure 6-4: Colour-shaded topographic map of the Piskanja area, showing SRK’s interpretation of potential faults that may affect the Piskanja project (drillholes are red circles). Faults away from the deposit are omitted

6.4.3 Slumping

The results of a pilot study completed to understand the potential for pre-lithification slumping throughout the regional borate-bearing sequence (illustrated in outcrop in Figure 6-5) have highlighted the potential presence of syn-sedimentary „or soft-sediment‟ faults as indicated in the Company‟s core photographs (Figure 6-6), and show the merit in extending this type of logging. Whilst SRK has not observed examples of slumped borates at Piskanja, there is no obvious geological reason why they should not be affected by slumping. If this is the case, then it could be expected that the borate distribution could be complex (folded and dismembered) on the scale of metres to tens of metres.

Given the significance of the small-scale geological continuity for mining, it is recommended that any borate intervals affected by, or directly flanked by, areas of sediment slumping are highlighted. Moreover, closer spaced infill drilling should be conducted in specific areas of the deposit, especially between holes with thicker intervals of borate, to help reduce the uncertainty in thickness variations at mine panel scale.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 17 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Figure 6-5: Chaotic soft-sediment deformation within a broadly concordant layer of claystones and sandstones, Bella Sten magnesite pit

Figure 6-6: Examples of pre-lithification structures, both extensional and contractional in nature, affecting sandstone and mudstones of the Piskanja Project

6.5 Mineralisation

6.5.1 Introduction

The Piskanja mineralisation is likely to have been deposited in a restricted inter-montane basin occupied by a perennial saline lake. Boron-rich fluids in these environments usually emanate from geothermal springs with a volcanic input (Garrett 1998).

The main mineralisation is on the whole concordant with stratigraphy and Erin‟s staff have noted that the borate mineralisation correlates laterally with carbonate horizons, consistent with a syn-depositional or syn-diagenetic origin anticipated for evaporitic deposits.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 18 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

6.5.2 Mineralisation Textures

The mineralised horizons show a range of textures comprising different growths of the borate minerals colemanite, ulexite and howlite predominantly. The main textures observed by SRK are summarised below and shown in Figure 6-7:

 Massive mineralisation which appears within pale brown, laminated carbonate rocks (Figure 6-7a). Occasional muddy irregular laminations and inclusions occur within the borates which appear to represent displaced soft-sediment (pre-lithification). These zones are interpreted to have developed at or just below the lake bed.

 Minor veinlets parallel to the stratigraphy (Figure 6-7b). These have mineral fibres oriented steeply, indicating the veins opened vertically, consistent with growth at low overburden pressures (very shallow depths). The veins have sub-angular tips, suggesting they developed when the sediments were only partially lithified (semi- coherent).

 Breccias hosted by siltstones and claystones, with textures ranging from clast-supported jigsaw breccias through to matrix-supported chaotic breccias with a fine clast size (<1mm to 2 cm; Figure 6-7c). These clast-size variations appear vertically stratified; tentatively suggesting variations in sediments (porosity and cohesion) influenced the style of deformation. Overall these breccias are interpreted to represent overpressuring by hydrothermal fluids in the shallow-subsurface.

 Both massive and vein borates contain two types of vuggy hollows;-

o Open vugs with ingrowing crystals of borate, which represent holes present during the growth or remobilisation of borate minerals (Figure 6-7d).

o Vugs which appear to have undergone some mineral dissolution, which are commonly stained by minor hydrocarbons (Figure 6-7e). These are interpreted as minor permeability networks where aggressive (acidic) fluids associated with the maturation of hydrocarbons have accentuated existing porosity. The hydrocarbons are likely to be locally sourced in organic rich units in the sediments.

Overall, there is very little textural evidence preserved for an active tectonic component controlling the ore, with most textures consistent with a syngenetic to diagenetic origin for the mineralisation. However, a subtle structural control is suggested by mineralisation thickness and dip (see below) and it may be worth mapping out the different styles of mineralisation to understand the controls better.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 19 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Figure 6-7: Mineralisation textures: (a) Remnant carbonate partings in massive borate; (b) Layer-parallel vein showing vertical opening direction; (c) Variably brecciated interval; (d) Primary vug; (e) Secondary dissolution vug

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 20 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Figure 6-8: Massive borate mineralisation in hole EVP2012-111 from 310.30 m to 313.20 m, situated at the contact between shale and dolomite units

6.5.3 Mineralisation Geometry

The thicker accumulations of borate mineralisation appear to occur within a broadly NW-SE orientated corridor some 200-250m wide, as illustrated for borate horizon (KZONE3) in Figure 6-9. This suggests a potential geological control on the deposition, such as faster subsidence due to faulting or the presence of a hydrothermal vent sourcing the B-bearing hydrothermal fluids, both of which point to a the presence of one or more faults. SRK tentatively interprets this corridor to be fault bound and believes there is some geological support for such an interpretation.

KZONE3 borate horizon wireframe

Preliminary fault interpretations

Corridor of increased mineralisation thickness?

Figure 6-9: Plan view of true thickness for borate horizon KZONE3

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 21 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

7 DEPOSIT TYPE

The Piskanja deposit is of continental lacustrine type, typical of many global boron deposits, and is considered to have formed within a closed basin with abnormally high salinity. The boron mineralisation is most likely to have been sourced from local volcanic rocks, from which it was leached by hydrothermal fluids. Boron minerals were then deposited in sedimentary successions in lacustrine conditions through the processes of evaporation and chemical precipitation. The presence of laminated dolomitic rocks and claystone in association with borate mineralisation indicates sedimentation in the deeper parts of a lake.

Most borate minerals are highly soluble in water which restricts the areas in which they form, and more importantly, are preserved. The majority of known global borate deposits have formed in lacustrine or playa lake environments in closed basins that opened up in active extensional setting near subductive plate boundaries. Rock types associated with the deposits generally include calc-alkaline extrusive rocks, tuff, limestone, marl, claystone, gypsum, continental silts and sands. The source of boron is not always the same and can be derived variously from leached marine sediments, magmatic fluids from subducted crust or from volcanic material (tuff).

The boron deposits in the USA and Turkey (which together account for some 80% of world production), are associated with continental sediments and show a continuum between hydrothermal spring, playa lake and lake deposits. Borate minerals precipitate once they become saturated in the fluids circulating these basins, either through evaporation of the basinal waters or addition of borate rich fluids from hydrothermal springs and circulating meteoric waters. Different borate minerals form at different levels of acidity; for example, borax (sodium borate) precipitates at a higher pH than ulexite, and in comparison colemanite forms at a lower pH and in warmer fluids. Due to cycles of basin refill and sediment input, there may be numerous layers of borate mineralisation interbedded with barren sedimentary horizons.

Borate deposits, due to their process of formation, are generally found as stratiform layers within basins, typically of Tertiary (Neogene) age and proximal to areas of volcanic activity of a similar age. Deposits showing these characteristics have already been identified and exploited in western Turkey at Kirka, Bigadiç and Kestelek among others. The origin of the borates within these deposits is related to mixing of borate-rich solutions within lacustrine basins controlled by evaporation (Helvaci and Alonso, 2000).

The Turkish deposits of Kirka, Bigadiç and Kestelek are owned and operated by Eti Maden. According to Eti Maden‟s website, (http://en.etimaden.gov.tr/) the Kirka deposit reportedly produces some 2.5 million tonnes per annum (tpa) of sodium borate ore at a mean grade of

26% B2O3. The Bigadiç deposit is reported to produce some 800,000 tpa of ulexite and

colemanite ore at between 29% and 31% B2O3. The Kestelek deposit produces some 200,000

tpa of colemanite ore with a mean grade of 29% B2O3 from an open pit.

The Rio Tinto owned Jadar project in northwest Serbia is a unique lithium borate deposit currently at a prefeasibility stage of evaluation with an Inferred Mineral Resource of 118Mt

containing 1.8% LiO2 and 16.2Mt B2O3 in the lower of three mineralised zones, (Rio Tinto 2012 Annual Report).

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 22 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Pan Global Resources has a number of joint venture properties in central Serbia at an early stage of exploration which it is exploring for deposits analogous to the Jadar deposit.

8 ERIN EXPLORATION

The historical exploration carried out in the region is summarised in Section 5 above. Exploration work completed and or managed by the Company since its involvement in 2010 has comprised:

 The collection and analysis of all publically available historical data relating to the Jarandol Basin and its geological setting, evolution, lithology, structure and mineralisation;

 Geological mapping (during 2012) of the Piskanja exploration licence and surrounding area at a scale of 1:5000. Due to the limited outcrop across the licence area, the Company was not able to undertake surface (soil or rock chip) sampling, trenching or pitting;

 Drilling and sampling as discussed in detail in Section 9 below, and

 Mineralogical studies on 47 singular and composite mineralised samples taken from a selection of the drillcore during December 2012. The studies involved petrographic, x-ray diffraction and scanning electron microscope with energy dispersive x-ray spectroscopy analysis of the main mineral phases, which concluded that the borate mineralogy at Piskanja is dominated by colemanite, ulexite and less commonly hydroboracite or jarandolite.

9 DRILLING 9.1 Introduction

This section summarises both the historical drilling undertaken before the Company‟s involvement and the drilling undertaken by the Company itself. The collar locations for all of this drilling are shown in Figure 9-1.

9.2 Ibar Mines (1987-1992)

Between 1987 and 1992, Ibar Mines completed 22 diamond core holes totalling 7,006 m of drilling to an average hole depth of 320 m. Total core recovery was reportedly very good (90- 100%) in shale, marl, sandstone and tuff horizons, but less so (60-75%) in volcanic breccia, breccia-conglomerate, conglomerate and borate mineralisation. A total of 89 core samples averaging 1 m in length were collected from 11 boreholes which intersected mineralisation all of which were analysed for boron. Mineralisation was identified in two horizons with an average thickness of 4.5 m for the upper bed and 3.5 m for the lower bed, lying between 50 m and 260 m depth.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 23 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

9.3 Ras Borati (1997)

During 1997, Ras Borati completed 10 reverse circulation (RC) holes, totalling 2,772 m. These holes were drilled by subcontractor Midnight Sun Drilling Co. Ltd, Canada, using a T685H Schramm drilling rig. A total of 204 chip samples were collected from 8 RC holes. The samples were prepared and analysed at the Geozavod-Nemetali laboratory in Belgrade using wet chemistry analysis.

9.4 Rio Tinto (2006-2007)

Rio Tinto acquired the Piskanja Project during 2006 as part of its regional investigation of borate potential in Tertiary basins across the Balkan region. An exploration drilling programme was completed, comprising a total of 16 holes for some 6,076 m of drilling. A total of 708 samples were prepared at the ITMNS laboratory in Belgrade and assayed by SGS Lakefield, Canada using potassium fusion ICP-AES as the primary method for determination of boron content.

9.5 Erin (2010- 2016)

9.5.1 Overview

The Company has undertaken two drilling programmes to date, with the latest phase of drilling and sampling completed during 2015.

The first phase of drilling undertaken by the Company was completed between July 2011 and December 2012. The programme comprised a total of 38 drillholes for 13,568 m of diamond core (“DC”) drilling and provided approximate 100x100m sample coverage across the deposit. A combination of the following Serbian drilling contractors was used to complete the drilling: GeoMag d.o.o, Silur d.o.o. and Geosonda d.o.o.

The latest phase of exploration drilling was a relatively small programme aimed at increasing the sample coverage to 50x50m within the central area of the deposit. DC drillholes for 2015 were collared on previously established drill section lines and drilled vertically. GeoMag d.o.o was retained to undertake the drilling.

In total a further 12 drillholes were drilled for 3,458 m of DC drilling. The positions of new drillhole collars for 2015 are illustrated in pink in Figure 9-1.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 24 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Preliminary fault interpretations

Mineralisation wireframes

Figure 9-1: Location of new collars (pink) completed by the Company during the 2015 exploration program

A total of 98 holes totalling some 32,880 m have now been completed at the Project. All drilling data available as of 31 May 2016 was made available to SRK. A summary of the completed drillholes is provided in Table 9-1 subdivided by company.

Table 9-1: Summary of Piskanja Drilling as at 31 May 2016 Company Date Count Total length (m) Ibar Mines 1987-1992 22 7,006.0 Erin Ventures through Ras borat 1997 10 2,771.9 RioTinto 2006-2007 16 6,076.0 Erin Ventures through Balkan Gold 2011-2012 38 13,568.1 Erin Ventures through Balkan Gold 2015 12 3,457.8 Total 98 32,880

9.5.2 Collar Surveys

The topographic survey of all the drillhole collars completed by the Company has been completed by total station using a Leica FexLine TS02.

9.5.3 Downhole Surveys

SRK has been supplied with downhole survey information for the start and the end of each hole completed by the Company, with intermediate readings at approximately every 10 m, collected using a digital probe by Geo-Log doo (Belgrade, Serbia). A measurement for deviation is provided every 2.5 mm is collected by the digital probe, the Company creates a composite reading for every 10m. In general, the data collected is considered to be of high precision and accuracy suitable for use in this resource estimation.

Historic drillholes were drilled at a vertical orientation. However, no downhole surveys were recorded for these holes.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 25 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

9.5.4 Hole Orientation

All drilling undertaken on the Project has been completed from surface at a vertical orientation. The drillholes are typically plotted on sections oriented E-W and NE-SW across the deposit and are spaced approximately 50–100 m apart, proving intersections at a similar spacing. Hole lengths range from 90 – 620 m and intersection angles with the mineralisation typically ranging from perpendicular to -45°.

It is SRK‟s view that the drilling orientations are reasonable to model the geology and mineralisation based on the current geological interpretation. Figure 9-2 provides a cross section to show the typical drilling orientation and dip of the mineralisation wireframe.

Figure 9-2: Example cross section through the Piskanja deposit

9.5.5 Diamond Drilling Procedure

The drilling was performed by contractors and managed by the Company‟s geological team. All drilling was completed using DC with double tube (Silur d.o.o) or triple tube (Geomag d.o.o) core barrels. Core was typically HQ in diameter (64mm).

Core was typically produced in 3 m core runs and the packed into plastic and metal core trays at the drill site. The core was subsequently washed, marked for down-hole direction and the core box marked with borehole ID and depth. An initial phase of geological logging was completed prior to transporting the core to the Company‟s office in Baljevac.

9.5.6 Core Recovery

SRK has reviewed the drill core recovery results and found that in general the recovery is good with an average recovery of 95% for the mineralised horizons (Figure 9-3) and 94% for the host rock.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 26 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Histogram for Core Recovery (%) at Piskanja 35

30

25

20

15 REC_PCT %Frequency

10

5

0 30 40 50 60 70 80 90 100 Core Recovery (%)

Figure 9-3: Core Recovery within mineralised horizons at the Piskanja Project

9.5.7 Core Storage

All diamond drill core completed by the Company is stored in a facility located in Baljevac, as shown in Figure 9-4.

Figure 9-4: Core logging and storage facility (2012)

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 27 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

9.6 SRK Comments

In the opinion of SRK, the sampling procedures used by the Company conform to industry best practices and the resultant drilling pattern, when combined with the historical holes, is sufficiently dense to interpret the geometry and boundaries of the borate mineralisation to a reasonable level of confidence.

10 SAMPLE PREPARATION, ANALYSIS AND SECURITY 10.1 Introduction

The following section relates primarily to the methods and protocols used by the Company during its exploration campaigns to date.

10.2 Diamond Drilling Sample Preparation and Chain of Custody

Core is transported from the drill site to the core storage facility where it is logged for geology and geotechnical parameters (i.e. core recovery and RQD) and digital photographs are taken. Sample lengths are then allocated guided by visually logged geological contacts and mineralisation styles (massive, intercalated or disseminated) and the core is subsequently split using a diamond core cutter. Half core samples are placed in to sample bags and numbered with a predefined sample number (Figure 10-1). Samples are transported to the preparation laboratory either by Company staff or by courier, with remaining core stored at the Company‟s facility in Baljevac.

Samples are checked-in at the preparation laboratory against a sample submission form, with subsequent dispatch to the analytical laboratory completed by DHL courier.

Figure 10-1: Sample bags prepared for transport to the laboratory

10.3 Sample Preparation and Analysis

Samples are submitted for preparation to the SGS Bor laboratory (Serbia), using the standard preparation procedure PRP86. The procedure comprises: drying the samples at 60°C for 8 hours; crushing to 1-2mm using a jaw crusher; selecting a 700 g split using a Jones riffle splitter; and, pulverisation to 75µm using a Labtech Essa LM5 mill.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 28 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

The following sample analytical procedures were used until January 2013 (or the end of the 2011/ 2012 drilling programme):

 SGS Lakefield (Canada) analysed the samples for (soluble) boron using aqua regia

digest („ARD‟) ICP-AES and volumetric titration for samples that exceeded 15% B2O3. A limited number of samples were also analysed for (total) boron by alkali (KOH) fusion ICP-AES. SGS Lakefield is ISO 17025 accredited;

 SGS Bor analysed a limited proportion (20%) of the samples by aqua regia digest ICP- MS.

After this date (or the start of the 2015 drilling program), samples were assayed for boron by

Na2O2-fusion ICP-AES at SGS Ankara, Turkey. SGS Ankara is ISO 17025 accredited.

Excluding the volumetric titration methodology, the Company receives analytical results from the Laboratory in the form of boron percent (%). During database compilation the Company

converts boron (B %) to borate (B2O3 %) using the following formula:

B2O3 % = [B % x 3.2199]

10.4 Specific Gravity Data

Samples collected for density determination comprise quarter core material from mineralised intervals. The samples vary in length from 9 to 25 cm and density is determined using the water immersion method where natural state (non-dried) samples are weighed in air, coated in paraffin and then weighed in water.

A total of 37 density measurements from mineralised material were supplied by the Company and the results of these are summarised below in Table 10-1.

Table 10-1: Summary of density statistics GROUP MEAN (g/cm3) MAX (g/cm3) MIN (g/cm3) Mineralisation 2.287 2.537 1.914

It is noticeable that there is a relatively significant variation in the density results, most likely due to the variations in the dominant borate minerals in each sample (for example, colemanite and ulexite have densities of 2.42 g/cm3 and 1.95 g/cm3, respectively), however this may also be due to intercalations of clay and dolomitic rock within the mineralised sample. It is also possible that variable water content in the samples adds to the variation in the density results.

SRK noted no clear relationship between density and B2O3% grade, however, and given the relatively limited number of samples elected to apply the average length-weighted density of 2.287 g/cm3 for the mineralisation domains in the block model for the purpose of the MRE presented here.

SRK recommends that additional density determinations are undertaken on the Company‟s drillcore and that this should include low grade samples, to increase the number of results available for analysis, further test for a relationship between density and grade and improve the confidence in the density model for the Project generally. SRK also recommends that any additional density sampling should record the weight of the sample following oven-drying prior to immersion in water, given the potential for moisture content to affect the density readings in the current database.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 29 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

11 DATA VERIFICATION 11.1 Introduction

The Company completes routine data verification as part of its on-going drilling programmes, comprising validation of sample results using both standards and blank samples which are inserted routinely into each batch submitted to the laboratory.

SRK notes that 467 m (58%) of the total 800 m of sampling inside mineralisation wireframes completed is supported by QAQC data, which relates to holes drilled following EVP2012-118 which was drilled during 2012.

The remaining 333 m (42%) of sampling inside the mineralisation wireframes is not supported by QAQC data, however forms part of the same mineralised body and (excluding the historic Rio Tinto drilling, 12%) underwent the same sample preparation and assay procedures at SGS Lakefield. These drillholes are interspersed with those that are supported by QAQC data, they are visually comparable with adjacent intersections with QAQC and also show comparable sample distributions and mean grades (Figure 11-1).

Figure 11-1: Composite sample grade histogram distributions for borate, showing data assayed with QAQC support (left) and without QAQC (right)

Additional verification work completed by the Company has comprised:-

 Re-logging of all remaining historical drillcore;

 Verification twin drilling (within 5m) of historic Ras Borati hole B-29/9 with EVP2011-100 (2011) in attempt to verify the results. SRK noted an approximate 5m vertical offset between the two main mineralised horizons intercepted in the drillholes, which showed similar mineralised thicknesses (albeit at slightly differing grades). Given the conflict in depth measurement, in context of the 1.5-5m average thickness of the mineralised horizons, SRK elected to exclude this hole for the purpose of grade interpolation.

Section 11.2 below presents an analysis of the available quality assurance and quality control (QAQC) data collected by the Company and Section 11.3 comments on additional verification work undertaken by SRK.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 30 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

11.2 Assay QA/QC

11.2.1 1987 - 2007

Whilst no routine QAQC procedures were in place during this period, SRK has:

 excluded the historical drilling completed by Ibar Mines and Ras Borati between 1987 and 1997 on the basis of poor data validation and no the uncertainty with regards this data has no relevance with regards the MRE presented here; and

 compared the results of the 2006/2007 Rio Tinto drillholes against more recent drilling completed by the Company and noted in general a reasonable comparison, as discussed above in Section 11.1.

11.2.2 2011 - 2012

Routine QAQC procedures were introduced during the 2012 drilling programme following drillhole EVP2012-118. This included the submission of blanks, standards and duplicates in every batch of samples, with an overall QAQC insertion rate for the period of 14%. QAQC

materials were analysed using a combination of titration, ARD and Na2O2-fusion.

Standards

The Company introduced 3 different standards into the analysis sample stream, which were developed for the project by Shea Clark Smith, Mineral Exploration Geochemistry (MEG), Nevada. The standards were based on material sourced from the JP PEU Resavica Pobrdje Borate Mine located some 2.6 km northwest of Piskanja, with statistical limits determined based on round robin analysis completed by the Company.

To date, the mean grades and standard deviations for the standards have not been externally certified.

Round robin analysis was completed at 8 separate laboratories using a combination of 2-acid

digest and Na2O2 fusion with ICP assay. A summary of the selected laboratories and assay results are presented in Appendix A to this report.

Based on a review of the round robin data (141 samples), SRK elected to exclude the following 31 results, which represents some 22% of the total round robin database for the purpose of deriving the MRE presented here. This included:

 the 2-acid digest analyses completed by Florin Analytical Services (Nevada), given the inconsistency noted in the primary laboratory (SGS Lakefield) assay results relating to acid digestion, and;

 the results from the Alex Stewart Laboratory (Argentina), given the indication for a low bias in the results when compared to the other round robin laboratories.

The mean and standard deviation values per standard for boron are shown in Table 11-1, with details relating to the accepted round robin results and summary statistics provided in Appendix A.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 31 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Table 11-1: Summary of Standard Material for boron submitted by the Company in sample submissions Boron; B (%) Standard Material Certified SD Company Value Low 1X B 6.17 0.25 Shea Clark Smith, Mineral Exploration and Mid 2X B 11.21 0.46 Geochemistry, Reno, USA High 3X B 14.5 0.59

SRK has reviewed the standard results for boron obtained using titration and is satisfied in general that they demonstrate (with the exception of a limited number of anomalies) a reasonable degree of accuracy at the assaying laboratory. With regards to sample submissions for relating to the use of aqua regia digest (ARD), SRK noted a reasonably significant bias in the medium and high grade standards towards higher grade (on average +20% for boron). SRK has accounted for the over-estimation of boron grade associated with ARD by applying a regression formula to the affected sample data.

The average results of the standard reference material submissions used in the QAQC programme to date are illustrated in Figure 11-2, with summary statistics and individual charts presented in Appendix A.

Blanks

A coarse marble blank from an outcrop at (approximately 40km by road northwest of Piskanja) was included in the sample stream, prior to sample preparation. Blank samples were inserted in to the sample stream associated with ARD at a rate of approximately 8%.

SRK has reviewed the results from the blank sample analysis, and (whilst there is an indication for low level presence of boron) has determined that in general there is little evidence for significant sample contamination at the preparation facility. The blank sample analysis chart is presented in Appendix A.

Duplicates

Field duplicates (from quarter core) were inserted into the routine sample submissions, at an overall rate of approximately 2%.

Whilst the quantity of data available for review is limited, the duplicates for boron analysed by titration (7 samples) show a good correlation to the original samples, with a correlation coefficient in excess of 0.9. The duplicates for boron analysed using ARD (13 samples) also show a reasonable correlation to the original samples, however with a slight indication for bias of the high-grade duplicates toward higher grade (on average +4%), which supports similar observations noted in the (ARD) QAQC standard results. Duplicate charts are presented in Appendix A.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 32 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Summary of titration boron assays at SGS Laboratories for Summary of ARD boron assays at SGS Laboratories for CRM CRM submissions vs CRM Limits submissions vs CRM Limits 18.00 18.00

16.00 16.00

14.00 Certified (%) 14.00 Certified (%) V_+2Stdv V_+2Stdv 12.00 V_-2Stdv 12.00 V_-2Stdv V_+3Stdv V_+3Stdv

10.00 V_-3Stdv V_-3Stdv Assay Assay B(%) Assay Assay B(%) 10.00 X=Y X=Y

8.00 Low 1X B 8.00 Low 1X B Mid 2X B Mid 2X B High 3X B 6.00 High 3X B 6.00

4.00 4.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 Certified Value B (%) Certified Value B (%) Figure 11-2: QAQC Standard Summary Charts from submission of Piskanja Samples (2011/2012) showing analysis by titration (left) and ARD (right)

Umpire Laboratory Analysis

A small number check samples were submitted to ALS Romania to verify the analytical performance at SGS Lakefield.

The duplicate data is presented in Appendix A. A comparison between the results analysed

by titration at SGS Lakefield and Na2O2-fusion at ALS Romania (11 samples, comprising coarse and pulp reject material) show a good correlation, with a coefficient in excess of 0.98. However, the results analysed using ARD at both the laboratories (30 samples) show a comparatively poor relationship (on average 18% lower at ALS Romania).

In addition, 6 low, 6 medium and 6 high grade standards were submitted to ALS Romania and

analysed using Na2O2-fusion. The results demonstrate a relatively good accuracy, which were (on average) within 1.5% of the accepted mean of the standards.

11.2.3 2015

All drilling captured inside mineralisation wireframes and completed during 2015 (253 m) is supported by QAQC data.

The QAQC system included the submission of blanks, standards and duplicates in every batch of samples submitted to SGS Ankara, with an overall QAQC insertion rate for the period

of 15%. All QAQC materials were analysed using Na2O2-fusion.

Standards

The Company has inserted 3 different standards into the analysis sample stream which were developed by MEG and round robin analysis, as summarised in Section 11.2.2. To date, the mean grades and standard deviations for the standards have not been externally certified.

SRK has reviewed the standard results for boron at SGS Ankara and (whilst in general these demonstrate a reasonable degree of precision), noted a slight bias toward lower grade, on average between 3-7% below the expected value. The average results of the standard reference material submissions used in the QAQC programme to date are illustrated in Figure 11-3, with summary statistics and individual charts presented in Appendix A.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 33 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Blanks

A coarse marble blank was inserted in to the sample stream at a rate of approximately 5%. SRK has reviewed the results from the blank sample analysis, and has determined that in general there is little evidence for sample contamination at the preparation facility. The blank sample analysis chart is presented in Figure 11-3.

Duplicates

Field duplicates (from quarter core) were inserted into the routine sample submissions, at an overall rate of approximately 5%. Excluding a single anomalous result, the duplicates for boron (38 samples) show a good correlation to the original samples, with a correlation coefficient in excess of 0.98. A duplicate chart is presented in Figure 11-3.

Verification Duplicates and Umpire Laboratory Analysis

Given the change in primary laboratory for the 2015 programme, the Company submitted 22 samples originally analysed at SGS Lakefield by titration to SGS Ankara and (as a control) to

BVM Perth for analysis using Na2O2-fusion. The comparison between BVM Perth and SGS Lakefield suggested a good overall correlation (with coefficient > 0.99), however the analysis at SGS Ankara suggested a slight bias towards high grade (on average 8%). Verification duplicate charts are presented in Appendix A.

In addition, during the 2015 analytical programme at SGS Ankara, the Company submitted 38

samples to BVM Perth for umpire analysis using Na2O2-fusion. A comparison between the results showed a good correlation between the samples, with a coefficient in excess of 0.99.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 34 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Summary of Na2O2-fusion boron assays at SGS Laboratories for CRM submissions vs CRM Limits 18.00

16.00

14.00 Certified (%) V_+2Stdv 12.00 V_-2Stdv V_+3Stdv V_-3Stdv

Assay Assay B(%) 10.00 X=Y

8.00 Low 1X B Mid 2X B

6.00 High 3X B

4.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 Certified Value B (%)

Analysis B Assays (%)-BLK Analysis B Assays (%)-Original Sample vs Field Duplicates

0.300 15 14

13 y = 0.9932x + 0.0956 0.250 12 R² = 0.9871

11

10 0.200 DL 9 X=Y

8 B SGS Ankara 3DL 0.150 7 B - SGS_Ankara (Anomaly Removed)

BDuplicate (%) Linear (B - SGS_Ankara (Anomaly

B Assay BAssay (%) 6 5DL Removed)) 5 0.100 4 SGS ANKARA 3

0.050 2

1

0 0.000 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 10 20 30 40 B Original (%) Assay Number Figure 11-3: QAQC Standard Summary Charts from submission of Piskanja Samples (2015) showing analysis for boron using Na2O2-fusion

11.3 Verifications by SRK

11.3.1 General verification

SRK has completed several visits to the Project during 2012, 2014 and 2016. During these visits SRK has reviewed drill cores for selected holes and the core logging and sampling procedures in place and discussed geological and structural interpretations and witnessed the extent of the exploration completed to date.

Prior to geological modelling, SRK also completed a data validation exercise on the digital sample database, the key findings of which are set out below.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 35 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

11.3.2 Sample Database

SRK noted a small number of sample interval overlaps and duplicate intervals in the database exports, which were rectified prior to data import. A subsequent check of the Company‟s assay database against the original laboratory certificates highlighted a number of data entry, sample ID sequence and boron to borate conversion errors which have also been corrected.

During the modelling process, SRK excluded the historic Ibar Mines and Ras Borati drillholes shown belowError! Reference source not found. that did not, in SRK‟s opinion, meet all spects of the validation procedure. These drillholes, which represent some 30% of the drilling database, were used as a guide for geological modelling but were excluded from the statistical and geostatistical analyses and grade interpolation process. The spatial position of the excluded drillholes and support provided by more recent drilling is illustrated in Figure 9-1.

Table 11-2: List of Drillholes Excluded from the 2016 MRE HoleID Company Comment B-10/91 Ibar Mines B-10/91??? Ibar Mines B-10/97 Erin Ventures through Ras borat B-11/91 Ibar Mines B-12/91 Ibar Mines B127/1 Ibar Mines B-13/91 Ibar Mines B-14/91 Ibar Mines B-15/91 Ibar Mines B-16/91 Ibar Mines B-16/97 Erin Ventures through Ras borat B-17/92 Ibar Mines Low confidence in depth measurements, which B-18/92 Ibar Mines results in up to a 20m offset from the borate B-19/92 Ibar Mines horizons intercepted in adjacent more recent drilling. B-2/89 Ibar Mines

B-20/92 Ibar Mines Only certain horizons appear to have been B-29/97 Erin Ventures through Ras borat sampled, based on highly selective sampling. B-3/90 Ibar Mines B-36/97 Erin Ventures through Ras borat There is no remaining historic drillcore, QAQC B-4/90 Ibar Mines or protocol information for drilling and sampling to help verify this phase of exploration. B-40/97 Erin Ventures through Ras borat B-41/97 Erin Ventures through Ras borat B-47/97 Erin Ventures through Ras borat B-48/97 Erin Ventures through Ras borat B-5/90 Ibar Mines B-6/90 Ibar Mines B-7/90 Ibar Mines B-8/91 Ibar Mines B-8/97 Erin Ventures through Ras borat B-9/91 Ibar Mines B-9/97 Erin Ventures through Ras borat B-B/91 Ibar Mines

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 36 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

11.3.3 Assay Technique

Based on a review of the 2011-2012 assay data, SRK noted a poor correlation between the results for boron derived using titration versus the results obtained using aqua regia digest („ARD‟), as illustrated in Figure 11-4 (170 samples), with the ARD methodology reporting on average 30% higher grades than titration. SRK was, however also able to compare the assay results for boron derived using titration with the results from some 82 samples that were also analysed by alkali fusion ICP-AES („fusion‟). This showed a good relationship, with a correlation coefficient in excess of 0.99 (Figure 11-5).

Given the above, the difference in the assay results between titration and the ARD methodology is considered by SRK to be most likely a result of inconsistent liberation of boron in acid digestion, related to boron‟s high susceptibility to volitisation from acid solution.

In order to account for the potentially significant over-estimation of boron grade associated with ARD, SRK has therefore given preference in the database to assay results derived by titration and fusion. For those samples where only the results from ARD are available (some 10% of the sample data inside the mineralisation wireframes), however, SRK has applied a regression formula to the affected sample data (as outlined below), based on the relationship from the graph shown in Figure 11-4:

B2O3 % Corrected = [B2O3 % from ARD – 3.5] / 1.14

B2O3% Titration vs. ARD 100

90

80

70 y = 1.1387x + 3.4689 R² = 0.7881 60

50 X=Y B2O3% B2O3% B2O3% ARD 40 Linear (B2O3%) 30

20

10

0 0 10 20 30 40 50 60 70 80 90 100 B2O3% Titration

Figure 11-4: Scatter plot for B2O3% samples (170) analysed by titration and ARD

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 37 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

B2O3% Titration vs. Fusion 60

y = 0.9982x + 0.5369 50 R² = 0.9917 X=Y

40

B2O3% Anomaly Excluded 30

B2O3% B2O3%Fusion 20

Linear (B2O3% Anomaly Excluded) 10

0 0 10 20 30 40 50 60 B2O3% Titration

Figure 11-5: Scatter plot for B2O3% samples (82) analysed by titration and fusion

11.3.4 Non-sampled Intervals

SRK noted the presence of a small number of non-sampled intervals for boron within the mineralised zones (some 1% of the database), which relate to visually weakly mineralised or non-mineralised intervals within the borate horizons and (in a limited number of cases) minor core loss. SRK has reviewed the non-sampled intervals on a case by case basis and treated the intervals as follows:

 non-sampled but weakly mineralised = replace with low grade (2.5% B2O3);

 non-sampled and not mineralised = replace with trace grade (0.01% B2O3);

 non-sampled associated with core loss = ignored during the sample compositing process i.e. treated as “missing”.

11.4 SRK Comments

SRK has completed a review of the available data and has rejected or corrected some of this for the purpose of grade interpolation. For the samples selected for use in the MRE, 42% of the data inside the mineralisation wireframes is not supported by QAQC, however these samples are well supported by, and interspersed with, more recent intersections which have acceptable QAQC results.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 38 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Whilst SRK has noted slight inconsistencies with regards to the QAQC assay results for 2015 which remain unexplained and require further investigation during future sampling programmes, SRK considers the overall sample preparation and laboratory performance at SGS Ankara to be appropriate and the data used to be suitable for the purpose of reporting the MRE at the level of confidence this has been reported to in this report.

For the holes drilled prior to EVP2012-119 (if available), the Company could also consider sending pulp splits from a representative portion of samples to the primary laboratory along with QAQC samples according to the current protocols to compare the laboratory performance today with its performance in 2006/2007 (for Rio Tinto drilling) and 2011/2012 prior to drillhole EVP2012-119.

12 MINERAL RESOURCE ESTIMATES 12.1 Introduction

The Mineral Resource model prepared by SRK and presented here utilises some 32,880 m of drilling for a total of 98 drillholes. The preparation of the MRE was supervised by Dr Mike Armitage. The effective date of the resource statement is 19 July, 2016.

The resource estimation methodology involved the following procedures:

 database compilation and verification;

 construction of wireframe geological models and definition of Resource domains;

 data conditioning (compositing and capping review) for statistical and geostatistical analysis;

 variography, block modelling and grade interpolation;

 resource classification and validation;

 assessment of “reasonable prospects for economic extraction” and selection of appropriate reporting cut-off grades; and

 preparation of the Mineral Resource Statement.

SRK was supplied with drilling data in a Microsoft Excel Database on 31 May 2016. The database was reviewed by SRK and imported into Datamine software.

SRK is satisfied with the quality of the database for use in the construction of the geological block model and associated MRE.

12.2 Statistical Analysis – Raw Data

An initial global statistical analysis was undertaken on the raw drill data. Summary statistics and incremental histograms were calculated. The distribution for borate is shown in Figure 12-1, with the separate populations noted in the assays relating to background host rock and low and high grade zones.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 39 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Figure 12-1: Incremental Histogram of Length Weighted Project Borate Assays

12.3 3D Modelling

12.3.1 Mineralisation Domains

SRK has modelled ten sub-parallel stratiform borate horizons to constrain the grade interpolation. While broadly similar to the geological interpretation used for the 2013 MRE, these have been updated to reflect the data now available and have been defined based on a

combination of lithological logging, borate grade (using an approximate 5% B2O3 cut-off) and visual assessments of geological and grade continuity.

Whilst SRK and the Company consider the borate horizons modelled to be an appropriate representation of the mineralisation at Piskanja, the definition of marker units in the stratigraphy (such as conglomerate horizons) would be a useful additional tool for adding confidence to the overall form and continuity given the stratiform nature of the mineralisation.

Preliminary fault interpretations (developed as part of the 2016 SRK structural study) were reviewed in context of the updated mineralisation domains. Given the indication for only slight changes in dip across the fault structures, no offsets have however been applied to the geological model.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 40 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

12.3.2 Statistical Analysis

The modelled domains were checked to ensure they formed appropriate sample populations for grade estimation with respect to borate grade, with the presence of any bimodal populations noted to ensure appropriate representation during block grade estimation. An example of the bimodal sample population for borate for the KZONE3 domain, which is a reflection of graduation from high to low grade observed both laterally and vertically as layering from top to bottom of the thicker horizons (related to mineralisation style), is illustrated in Figure 12-2.

The presence of higher and lower grade areas and vertical grade layering is shown for the KZONE3 domain Figure 12-3 and Figure 12-4 respectively.

Figure 12-2: Log histogram plot for borate for domain KZONE3

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 41 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Borate horizon KZONE3 sample KZONE3 composites Higher grade borate mineralisation

Lower grade borate mineralisation

Figure 12-3: Assessment of 3D borate grade distribution within the KZONE3 domain

Local borate grade layering

within thicker horizons

EVP2015

-

141

EVP2015

EVP2015

-

139

- 145

Figure 12-4: 2D assessment of borate grade distribution downhole, looking west

12.3.3 Mineralisation Model Coding

Domain names and a summary description of the modelled wireframes are given in Table 12-1 while Figure 12-5 and Figure 12-6 provide example illustrations of these. While the wireframes were developed by SRK they have been discussed, and agreed, with the Company.

The mineralisation modelled comprises ten separate horizons which are geologically continuous along strike for between 250m and 700 m, which have dip extents of up to 1 km and have average thicknesses typically between 1.5 and 5 m, reaching up to 20 m in certain areas.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 42 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Table 12-1: Summary of Mineralisation Zones at the Piskanja Project GROUP KZONE Wireframe Deposit Description Borate mineralised Borate horizons hosted by a series of alternating 100 1-10 horizons Piskanja units of mudstone, siltstone and sandstone (ob1_tr - ob10_tr)

Borate horizons (KZONE 1-10)

Preliminary NW-SE fault interpretations

Figure 12-5: Piskanja Mineralisation Model, looking north

Borate horizons (KZONE 1-10)

Preliminary fault interpretations

Figure 12-6: Piskanja Mineralisation Model, looking down

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 43 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

12.4 Compositing

Within the thicker zones of mineralisation, the borate grade data shows the presence of vertical layering from hangingwall to footwall. Given this, SRK elected to create 1m composites to ensure sufficient resolution during block grade interpolation whilst honouring the mean sample length within the mineralised zones.

12.5 Evaluation of Outliers

SRK completed an analysis for outliers based on a review of the statistical distribution of the assay values in each domain and a review of where higher grade samples were located. Based on this no high-grade capping was applied.

Raw and log histograms plots (as illustrated for borate composites for KZONE3 in Figure 12-7) are shown per mineralisation (KZONE) domain in Appendix B. Table 12-2 provides a summary of the composite sample statistics for each of the estimation (KZONE) domains.

Figure 12-7: Raw and Log Histograms for borate for the KZONE3 domain

Table 12-2: Composite Statistics for Borate KZONE FIELD NSAMP MIN MAX MEAN VAR STDDEV COV 1 B2O3_PCT 244 0.01 55.38 34.05 178.43 13.36 0.39 2 B2O3_PCT 86 0.01 67.30 31.57 238.24 15.43 0.49 3 B2O3_PCT 264 0.01 53.45 31.68 211.85 14.56 0.46 4 B2O3_PCT 66 0.01 53.45 35.79 185.69 13.63 0.38 5 B2O3_PCT 11 2.58 42.83 28.92 188.73 13.74 0.48 6 B2O3_PCT 124 0.01 45.07 16.59 187.80 13.70 0.83 7 B2O3_PCT 21 0.01 40.14 15.37 170.43 13.05 0.85 8 B2O3_PCT 9 10.12 32.12 20.65 54.70 7.40 0.36 9 B2O3_PCT 9 5.38 53.13 32.60 349.84 18.70 0.57 10 B2O3_PCT 10 2.50 23.87 13.36 50.24 7.09 0.53

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 44 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

12.6 Geostatistical Analysis

SRK elected to treat the mineralisation domains as a single zone for variography purposes. This increased the number of available sample pairs for analysis and honoured the likely close genetic relationship between the different domains. The experimental variogram was calculated using a 20m off-plane tolerance in attempt to reduce the influence of sample data from spatially separate horizons from impacting the assessment of grade continuity.

Experimental semi-variograms were calculated in the along-strike, down-dip and across-strike orientations, with a short-lag down-hole variogram calculated to characterise the nugget effect. Directional variograms were in general poorly defined and therefore omni-directional structures were selected for fitting of the final variogram models. All variances were re-scaled for each mineralised zone to match the total variance (”VAR”) for that zone.

The variogram model and parameters for the combined Mineralisation domain GROUP 100 for borate is shown in Table 12-3.

Table 12-3: Modelled semi-variogram parameters for Domain (GROUP 100)* Variogram Parameter B2O3_PCT-GROUP100 Co 35.00 C1 149.71 A1 – Along Strike (m) 40 A1 – Down Dip (m) 40 A1 – Across Strike (m) 40 C2 50.60 A2 – Along Strike (m) 265 A2 – Down Dip (m) 265 A2 – Across Strike (m) 265 C3 0.00 A3 – Along Strike (m) 0 A3 – Down Dip (m) 0

A3 – Across Strike (m) 0 Nugget Effect (%) 15%

* Variogram structures are subsequently re-scaled to the total sample variance per estimation KZONE

12.7 Block Model and Grade Interpolation

A block model prototype was created based on Serbian Gauss Kruger - Zone 7 coordinate system. Block model parameters were chosen to appropriately reflect the grade variability both laterally and vertically from hangingwall to footwall.

SRK notes that while the selected block size is small relative to the average drillhole spacing in the less well drilled areas of the deposit, SRK used sufficiently expanded search ellipses and elevated sample numbers when interpolating block grades in areas away from the closer spaced drilling.

To improve the geometric representation of the geological model, sub-blocking was allowed along the boundaries to a minimum of 5x5x0.5 m (x, y, and z). A summary of the block model parameters is given in Table 12-4.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 45 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Table 12-4: Block Model Dimensions Number of Min Sub- Model Dimension Origin (UTM) Block Size Blocks blocking (m) X 7471410 10 120 5 Piskanja Y 4803610 10 84 5 Z -186 2 300 0.5

12.8 Final Interpolation Parameters

Ordinary Kriging (“OK”) was used for the grade interpolation. Search ellipses were orientated to follow the trend of each domain using Datamine‟s Dynamic Anisotropy and domain boundaries were treated as hard boundaries during the interpolation process.

Inverse distance weighting squared (“IDW2”) was used for verification of the OK estimates. SRK applied the average length-weighted borate sample density of 2.287 g/cm3 to the block model.

The selected estimation parameters have been verified based on the results of a quantitative Kriging Neighbourhood Analysis (“QKNA”), and are presented in Table 12-5.

Table 12-5: Final Interpolation Parameters Estimation Parameters Description KZONE 1 – 10 Kriging zones for estimation FIELD B2O3_PCT Field for interpolation SREFNUM 1 Search reference number SMETHOD 2 Search volume shape (2 = ellipse) SDIST1 65 Search distance 1 (dip) SDIST2 65 Search distance 2 (strike) SDIST3 10 Search distance 3 (across strike) SANGLE1 Dynamic Search angle 1 (dip direction) SANGLE2 Dynamic Search angle 2 (dip) SANGLE3 0 Search angle 3 (plunge) SAXIS1 3 Search axis 1 (z) SAXIS2 1 Search axis 2 (x) SAXIS3 3 Search axis 3 (z) MINNUM1 6 Minimum sample number (SVOL1) MAXNUM1 20 Maximum sample number (SVOL1) SVOLFAC2 2 Search distance expansion (SVOL2) MINNUM2 6 Minimum sample number (SVOL2) MAXNUM2 20 Maximum sample number (SVOL2) SVOLFAC3 4 Search distance expansion (SVOL3) MINNUM3 2 Minimum sample number (SVOL3) MAXNUM3 20 Maximum sample number (SVOL3) MAXKEY 2 Maximum number of samples per drillhole SANGL1_F TRDIPDIR Dynamic Anisotropy SANGL2_F TRDIP Dynamic Anisotropy

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 46 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

12.9 Model Validation and Sensitivity

12.9.1 Sensitivity Analysis

Grade Interpolation was performed using Datamine, based on optimum parameters verified through the QKNA exercise. The exercise was based on varying kriging parameters for borate (namely number of samples and search ellipse size) to reflect a number of different scenarios. Testwork was focused on the KZONE3 domain given its significant contribution to metal (22%) in the geological model and typical representation of the deposit as a whole in terms of drillhole spacing and borate grade distribution.

Whilst SRK noted a limited degree of sensitivity in the mean block grade to changes in the interpolation parameters, block grades (visually) better reflected the overall grade distributions shown by sample composites by restricting the search ellipse dimension and maximum number of composites per drillhole to within reasonable limits, with the associated sensitivity shown in Appendix B. The final parameters were selected to ensure that the spatial grade variability within the deposit, both laterally and vertically as layering from hangingwall to footwall, was appropriately reflected in block grade estimates.

12.9.2 Block Model Validation

SRK has validated the block model using the following techniques:

 visual inspection of block grades in comparison with drillhole data;

 sectional validation of the mean samples grades in comparison to the mean model grades; and

 comparison of block model statistics.

Visual Validation

Visual validation provided a comparison of the interpolated block model on a local scale. A thorough visual inspection was undertaken in section and 3D, comparing the sample grades with the block grades. This demonstrated a good comparison between local block estimates and nearby samples, without excessive smoothing in the block model. Figure 12-8 to Figure 12-9 provide examples of these visual validation checks.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 47 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Borate horizons (drillhole Borate horizons intercepts) (block model)

Figure 12-8: Piskanja Block Model Borate Grade Distribution (3D view, looking north)

116

-

142

140

-

-

EVP2012 EVP2015 EVP2015 Figure 12-9: Piskanja Block Model Borate Grade Distribution (cross-section)

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 48 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Sectional Validation

As part of the validation process, the input composite samples were compared to the block model grades within a series of coordinates (based on the principle directions). The results of this were then displayed on charts to check for visual discrepancies between grades. Figure 12-10 shows the results for the borate grades for the domain KZONE1 based on section lines cut along x-coordinates.

The resultant plots show a reasonable correlation between the block model grades and the composite grades, with the block model showing a typically smoothed profile of the composite grades as expected. SRK notes that in less densely sampled areas, minor grade discrepancies do exist on a local scale. Overall, however, SRK is confident that the interpolated grades reflect the available input sample data and the estimate shows no sign of material bias. Validation plots for all domains are shown in Appendix B.

Statistical Validation

The mean block estimate for each domain was compared to the mean of the composite samples in each case (Table 12-6). The percentage difference varied between 0 and 7%, which SRK deems to be within acceptable levels.

SRK notes a slightly higher percentage difference in the means for mineralisation domains KZONE 7, 9 and 10 which is as a result of the sample means in these cases being skewed by a few high grade samples that influence a relatively small proportion of the block model tonnage. Comparison between the means for KZONE 9 and 10 are better for OK than IDW2 which is presumably due to the kriging applying a more appropriate level of smoothing (defined by the short-lag variogram) for the local block grade estimates.

Based on the visual, sectional and statistical validation results SRK has accepted the grades in the block model.

50 80

45 70 40 60 35 50 30

25 40 B2O5%

20 No.Samples 30 15 20 10 10 5

0 0 7471400 7471600 7471800 7472000 7472200 7472400 7472600 X-Co-ordinate

Sample Mean Model Mean No. Samples

Figure 12-10: Validation Plot (Easting) showing Block Model Estimates versus Sample Mean (40m Intervals) for domain KZONE1 for borate

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 49 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Table 12-6: Summary Block Statistics for Ordinary Kriging and Inverse Distance Weighting Interpolation Methods Block Absolute Estimation Composite KZONE Field Estimate % Difference Difference Method Mean (%) Mean (%) (%) OK 33.04 34.05 -3.0% -1.01 1 B2O3 IDW 32.65 34.05 -4.1% -1.40 OK 29.62 31.57 -6.2% -1.95 2 B2O3 IDW 29.45 31.57 -6.7% -2.12 OK 30.62 31.68 -3.4% -1.06 3 B2O3 IDW 30.67 31.68 -3.2% -1.00 OK 34.60 35.79 -3.3% -1.19 4 B2O3 IDW 33.76 35.79 -5.7% -2.03 OK 31.02 28.92 7.3% 2.11 5 B2O3 IDW 31.14 28.92 7.7% 2.22 OK 17.69 16.59 6.6% 1.10 6 B2O3 IDW 16.47 16.59 -0.7% -0.12 OK 12.19 15.37 -20.7% -3.18 7 B2O3 IDW 14.40 15.37 -6.3% -0.97 OK 22.06 20.65 6.8% 1.41 8 B2O3 IDW 20.66 20.65 0.0% 0.01 OK 32.78 32.60 0.6% 0.18 9 B2O3 IDW 29.11 32.60 -10.7% -3.49 OK 13.01 13.36 -2.6% -0.35 10 B2O3 IDW 11.13 13.36 -16.7% -2.23

12.10 Mineral Resource Classification

Block model quantities and grade estimates for the deposit were classified according to CIM Standards.

Mineral Resource classification is typically a subjective concept, industry best practices suggest that resource classification should consider both the confidence in the geological continuity of the mineralised structures, the quality and quantity of exploration data supporting the estimates and the geostatistical confidence in the tonnage and grade estimates. Appropriate classification criteria should aim at integrating both concepts to delineate regular areas at similar resource classification.

Data quality, geological confidence, sample spacing and the interpreted continuity of grades controlled by the deposit has allowed SRK to classify the block model in the Indicated and Inferred Mineral Resource categories. The following guidelines apply to SRK‟s classification:

Measured

No Measured Mineral Resources have been reported due to the additional work that needs to be completed by the Company to improve confidence in the data quality of the sample database (as noted in Section 14). In addition, uncertainty in the geological and grade continuity at a mining scale will need to be addressed thorough further infill drilling.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 50 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Indicated

Indicated Mineral Resources comprise the blocks which show a good level of geological confidence and grade continuity and which are situated within relatively well drilled areas of the model and typically up to 75 m beyond these areas.

Inferred

Inferred Mineral Resources are in domains that display reasonable geological confidence and where blocks are typically within 100 m of sample data. These areas require infill drilling to improve the quality of the geological interpretation and local block grade estimates to a level suitable for mine planning.

An example of SRK‟s Mineral Resource classification for the Piskanja deposit is shown in Figure 12-11.

Zones removed from Resource (grey) to honour minimum mining thickness (1.2m)

Figure 12-11: SRK Mineral Resource Classification

12.11 Mineral Resource Statement

The CIM Standards require Mineral Resources to have reasonable prospects for economic extraction. In 2014 SRK completed a Preliminary Economic Assessment (PEA) which demonstrated the mineralisation at Piskanja to be potentially amenable to underground mining and based on the technical work and economic analysis presented in that report has reported all of the mineralisation outlined in producing this update which is above a cut-off of

12% B2 O3 over a minimum mining thickness of 1.2 m as a Mineral Resource.

SRK‟s Mineral Resource Statement for the Piskanja deposit derived in this manner is shown in Table 12-7.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 51 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Table 12-7: SRK Mineral Resource Statement as at 19 July 2016 for the Piskanja Deposit prepared in accordance with the CIM Standards

Category Cut-off Tonnes Mt B2O3 Grade % B2O3 Mt Indicated 7.8 31.0 2.4 12% B2O3 Inferred 3.4 28.6 1.0

12.12 Grade Sensitivity Analysis

The results of grade sensitivity analysis are shown in Table 12-8. This shows the Mineral Resource to be relatively insensitive to changes in cut-off. The tonnages and grades in these tables, however, should not be interpreted as Mineral Resources.

Table 12-8: Gradations for Open Pit Material at Piskanja at various Borate % Cut-off Grades Grade - Tonnage Table, Piskanja 19 July 2016 Indicated Inferred Cut-off Grade Quantity Borate Quantity Borate Borate (%) (Mt) Grade (%) Metal (Mt) (Mt) Grade (%) Metal (Mt) 5 8.3 29.6 2.5 3.7 27.2 1.0 6 8.3 29.7 2.5 3.6 27.4 1.0 7 8.3 29.7 2.5 3.6 27.5 1.0 8 8.2 30.0 2.4 3.6 27.8 1.0 9 8.1 30.2 2.4 3.5 28.1 1.0 10 8.0 30.5 2.4 3.5 28.3 1.0 11 7.9 30.7 2.4 3.4 28.5 1.0 12 7.8 31.0 2.4 3.4 28.6 1.0 13 7.7 31.2 2.4 3.4 28.8 1.0 14 7.6 31.4 2.4 3.3 29.0 1.0 15 7.5 31.6 2.4 3.3 29.2 1.0 16 7.4 31.9 2.4 3.2 29.4 0.9 17 7.2 32.2 2.3 3.1 29.7 0.9 18 7.1 32.5 2.3 3.1 29.9 0.9

12.13 Comparison to Previous Mineral Resource Estimates

In comparison to the previous SRK 2013 MRE for the Project, which was also reported at a

cut-off grade of 12% B2O3 but above a minimum mining thickness of 1.0 m, this updated MRE (which is reported above a minimum mining thickness of 1.2 m) has more borate in the Indicated category (2.4Mt compared to 1.7 Mt) and less borate in the Inferred category (1.0Mt compared to 1.8 Mt). These changes are primarily due to some previously reported Inferred material being upgraded in to the Indicated category though there is also a slight reduction in

B2O3 content for the Project as a whole from 3.5 Mt to 3.4 Mt (-4%) mainly as a result of a 6% reduction in tonnage. This change is primarily because the infill drilling has shown some of the thicker zones to be less continuous than previously assumed but also enabled the higher grade zones to be better delineated, which has reduced the tonnage slightly but at the same time slightly improved the grade (namely a global 2% increase in grade compared to the previous estimate).

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 52 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

SRK considers that the key changes in the MRE result from a combination of the following factors:

 improved confidence in the overall geological continuity at the deposit peripheries;

 improved confidence in the overall data quality through the verifying and (where required) amending transcription errors based on the original laboratory assay certificates;

 new infill drilling having changed the interpretation of some of the previously assumed thicker zones of mineralisation;

 more detailed grade modelling, the new infill drilling having improved the definition and continuity of grade layering which has improved the grade overall; and

 increasing of the minimum mining width from 1.0 m to 1.2 m, based on the work completed for SRK‟s 2014 PEA which was produced subsequent to SRK‟s previous MRE.

The application of a regression formula to the sample data analysed by ARD and replacement of non-sampled intervals with low grade values will have also had a relatively minor impact on the results.

13 INTERPRETATIONS AND CONCLUSIONS

The infill drilling and database verification work completed since the last MRE was produced has added further confidence to the geological model and borate grade distributions and enabled the production of an updated MRE which is similar in quantum, but is now more robust, than that previously produced.

There are, however, still areas of lower geological confidence which require more drilling and which may be subject to further revision in the future and there is also some associated work which could be done alongside or prior to more drilling to improve confidence in the estimate generally. This work is set out below.

14 RECOMMENDATIONS

Given the overall relatively small change in the Mineral Resource tonnage and grade, SRK recommends that further work should now focus on improving confidence in the geological and grade continuity through infill drilling at a mining scale, namely 10 or 20 m, in selected key areas of the deposit.

Improvements to data quality could also be achieved by:

 Completing additional density determinations to increase the number of results available for analysis and the confidence in the density model for the Project. SRK strongly recommends that any additional density sampling should record the weight of the sample following oven-drying prior to immersion in water, given the potential for moisture content to affect the density readings in the current database;

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 53 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

 Re-assaying by fusion of the following low-grade samples within the mineralised zones, which appear anomalously low compared with the adjacent drillhole intercepts: sample 2506 from EVP2015-137 and samples 2355 to 2356 from EVP2011-108;

 Submitting the same 20 samples sent to SGS Lakefield (titration) and SGS Ankara (fusion) to another certified laboratory (for assay by fusion) to further investigate the bias (+8%) noted in the verification duplicates at SGS Ankara;

 Completing an additional phase of Round Robin analysis on the standard materials to further reduce uncertainty in the expected values and standard deviations and obtain external industry certification for the analytical results;

 Rectifying the sample interval overlaps and sample sequence, transcription and B% to

B2O3% conversion errors in the Company‟s assay database to reduce the need to re- validate the assays during future MRE updates. Furthermore, SRK recommends that the Company compiles a single clean interval table file in which to store the assays for boron, with separate columns reflecting different assay methodologies and a „final‟ assay field to ensure an easily-auditable database. Acquiring a commercial database storage system would assist with ensuring that the database is fully validated;

 Completing further work to fully understand the extent of sub-horizontal zones of faulting affecting the upper borate horizons (KZONE3 and KZONE4) and, significantly, the geotechnical implications of these for mining of these horizons; and

 Completing a drillhole re-logging exercise to highlight any borate intervals affected by or directly flanked by areas of sediment slumping, given the significance of this in terms of the short-scale geological continuity for mining, and also to incorporate the defining marker units in the stratigraphy (such as conglomerate horizons), given their use in adding confidence to the overall form and continuity of the stratiform mineralisation.

15 REFERENCES

Andrić, N, Fügenschuh, B., Životić, D., Cvetković, V. 2015. The thermal history of the Miocene Ibar Basin (Southern Serbia): new constraints from apatite and zircon fission track and vitrinite reflectance data. Geologica Carpathica. 66, p. 37–50.

Garrett, D.E. 1998, Borates: Handbook of deposits, processing, properties, and use. Academic Press, San-Diego, 483 p.

Geosystem srl, Magnetotelluric Survey, Jarandol Basin, Serbia, 2006,

Ilic, M., Eric, V., 2009, Final Report on Exploration for period from 1st Sep 2006 to 21st July 2009, submitted on 27 July 2009 to Serbian Ministry for Environment, Mining and Spatial Planning. GEOEXPLORER PROJECT DOO on behalf Rio Sava Exploration doo. Belgrade. (in Serbian)

Lakefield Research limited, Lakefield, Ontario, Canada., 1998, A Quantitative Determination of The Boron Content of Borate Samples from Ras Borati, submitted by Erin Ventures Inc.

Matenco, L. & Radivojević, D. 2012. On the formation and evolution of the Pannonian Basin:Constraints derived from the structure of the junction area between the Carpathians and Dinarides. Tectonics 31, p. 1-31.

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 54 of 55 SRK Consulting Piskanja MRE 2016 – Main Report

Obradovic, J.Stamatakis, M.G., Anicic, S., Economou, G.S. 1992. Borate and borosilicate deposits in the Miocene Jarandol Basin, Serbia, Yugoslavia. Economic Geology, 87, p. 2169- 2174.

SGS Minerals Services, 2012, Report on magnetic and HTE testing of borate samples from Serbia,

Stojanovich, D., 1967, Howlite from Jarandol Tertiary basin, Proceedings from VI Congress of Geology in Ohrid, Macedonia.

University of Belgrade, Faculty of Mining and Geology, 2012, Testing of samples from the Piskanja borate deposit (translation from Serbian)

University of Belgrade, Faculty of Mining and Geology, 2012, Petrological characteristics of holes 104, 105, 106, 107, IBM-4 and IBM-6 – Piskanja, (in Serbian)

University of Belgrade, Faculty of Mining and Geology, 2013, Study of engineering properties rock masses and terrains of the Piskanja borate deposit (translation of concluding remarks from Serbian)

SRK Consulting UK Ltd., 2014, NI43-101 Technical Report and Preliminary Economic Assessment for the Piskanja Borate Project, Serbia

SRK Consulting UK Ltd., 2016, An Evaluation of the Structural Geology of the Piskanja Borate Project, Serbia

For and on behalf of SRK Consulting (UK) Limited

Mike Armitage, Chairman and Corporate Consultant, Project Manager and Project Director SRK Consulting (UK) Limited

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page 55 of 55 SRK Consulting Piskanja MRE 2016 – Technical Appendix A

APPENDIX

A QAQC ANALYSIS

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page A1 of A13 SRK Consulting Piskanja MRE 2016 – Technical Appendix A

STANDARDS – ROUND ROBIN ANALYSIS

Accreditation of laboratories used in the round robin analysis used to determine standard reference materials for the Piskanja Project (2013)

Laboratory Name Location Accreditation Analytical Method used

American Assay Sparks, Nevada, USA ISO/IEC 17025:2005 Na2O2-fusion ICP-AES

ACME Analytical Vancouver, Canada ISO/IEC 17025:2005 Na O -fusion ICP-AES Laboratories 2 2

Activation Laboratories Ontario, Canada ISO/IEC 17025 and CAN-P-1579 Na2O2-fusion ICP

ALS Chemex Vancouver, Canada ISO/IEC 17025:2005 Na2O2-fusion ICP-AES

Application pending (ISO/IEC Florin Analytical Services Reno, Nevada, USA Na O -fusion ICP 17025) 2 2

Genalysis/Intertek Perth, Australia ISO/IEC 17025 Na2O2-fusion ICP-AES

Application pending (ISO/IEC 17025)

Kalassay Perth, Australia Accredited by the National Na2O2-fusion ICP-AES Association of Testing Authorities, Australia

ISO 9001:2008 ISO 14001:2004 Alex Stewart Mendoza, Argentina ISO 17025:2005 (only for Au by fire Na2O2-fusion -fusion ICP assay and Li and K liquid brine analysis)

Round Robin Results – Raw Data (2013)

Low 1X B

7.1 AMERICAN ASSAY - Na2O2-fusion ICP-AES 6.9 Mean

6.7 V_+2Stdv

V_-2Stdv 6.5 V_+3Stdv

6.3 V_-3Stdv

6.1 ACME-VANCOUVER - Na2O2-fusion ICP-AES ACTIVATION LABS - Na2O2-fusion ICP 5.9 ALS-VANCOUVER - Na2O2-fusion ICP-AES

5.7 FLORIN ANALYTICAL - Na2O2-fusion ICP

5.5 FLORIN ANALYTICAL 2A - 2-Acid ICP GENALYSIS / INTERTEK - Na2O2-fusion ICP-AES 5.3 KALASSAY - INSPECTORATE - Na2O2-fusion ICP-MS 5.1 ALEX STEWART-MENDOZA - Na2O2-fusion ICP 0 10 20 30 40 50

MEAN 6.11 STD DEV 0.30 COUNT 46

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page A2 of A13 SRK Consulting Piskanja MRE 2016 – Technical Appendix A

Mid 2X B

13 AMERICAN ASSAY - Na2O2-fusion ICP-AES

12.5 Mean

V_+2Stdv 12 V_-2Stdv

V_+3Stdv 11.5 V_-3Stdv 11 ACME-VANCOUVER - Na2O2-fusion ICP-AES ACTIVATION LABS - Na2O2-fusion ICP

10.5 ALS-VANCOUVER - Na2O2-fusion ICP-AES

FLORIN ANALYTICAL - Na2O2-fusion ICP 10 FLORIN ANALYTICAL 2A - 2-Acid ICP

GENALYSIS / INTERTEK - Na2O2-fusion ICP-AES 9.5 KALASSAY - INSPECTORATE - Na2O2-fusion ICP-MS 9 ALEX STEWART-MENDOZA - Na2O2-fusion ICP 0 10 20 30 40 50

MEAN 11.07 STD DEV 0.58 COUNT 48

High 3X B

17 AMERICAN ASSAY - Na2O2-fusion ICP-AES 16.5 Mean

16 V_+2Stdv

V_-2Stdv 15.5 V_+3Stdv

15 V_-3Stdv

14.5 ACME-VANCOUVER - Na2O2-fusion ICP-AES ACTIVATION LABS - Na2O2-fusion ICP 14 ALS-VANCOUVER - Na2O2-fusion ICP-AES

13.5 FLORIN ANALYTICAL - Na2O2-fusion ICP

13 FLORIN ANALYTICAL 2A - 2-Acid ICP GENALYSIS / INTERTEK - Na2O2-fusion ICP-AES 12.5 KALASSAY - INSPECTORATE - Na2O2-fusion ICP-MS 12 ALEX STEWART-MENDOZA - Na2O2-fusion ICP 0 10 20 30 40 50

MEAN 14.34 STD DEV 0.71 COUNT 47

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page A3 of A13 SRK Consulting Piskanja MRE 2016 – Technical Appendix A

Round Robin Results – Accepted Data (2013)

Low 1X B

7.1 AMERICAN ASSAY - Na2O2-fusion ICP-AES 6.9 Mean

6.7 V_+2Stdv

V_-2Stdv 6.5 V_+3Stdv

6.3 V_-3Stdv

6.1 ACME-VANCOUVER - Na2O2-fusion ICP-AES ACTIVATION LABS - Na2O2-fusion ICP 5.9 ALS-VANCOUVER - Na2O2-fusion ICP-AES

5.7 FLORIN ANALYTICAL - Na2O2-fusion ICP

5.5 FLORIN ANALYTICAL 2A - 2-Acid ICP GENALYSIS / INTERTEK - Na2O2-fusion ICP-AES 5.3 KALASSAY - INSPECTORATE - Na2O2-fusion ICP-MS 5.1 ALEX STEWART-MENDOZA - Na2O2-fusion ICP 0 10 20 30 40 50

MEAN 6.17 STD DEV 0.25 COUNT 36

Mid 2X B

13 AMERICAN ASSAY - Na2O2-fusion ICP-AES

12.5 Mean

V_+2Stdv 12 V_-2Stdv

V_+3Stdv 11.5 V_-3Stdv 11 ACME-VANCOUVER - Na2O2-fusion ICP-AES ACTIVATION LABS - Na2O2-fusion ICP

10.5 ALS-VANCOUVER - Na2O2-fusion ICP-AES

FLORIN ANALYTICAL - Na2O2-fusion ICP 10 FLORIN ANALYTICAL 2A - 2-Acid ICP

GENALYSIS / INTERTEK - Na2O2-fusion ICP-AES 9.5 KALASSAY - INSPECTORATE - Na2O2-fusion ICP-MS 9 ALEX STEWART-MENDOZA - Na2O2-fusion ICP 0 10 20 30 40 50

MEAN 11.21 STD DEV 0.46 COUNT 38

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page A4 of A13 SRK Consulting Piskanja MRE 2016 – Technical Appendix A

High 3X B

17 AMERICAN ASSAY - Na2O2-fusion ICP-AES 16.5 Mean

16 V_+2Stdv

V_-2Stdv 15.5 V_+3Stdv

15 V_-3Stdv 14.5 ACME-VANCOUVER - Na2O2-fusion ICP-AES ACTIVATION LABS - Na2O2-fusion ICP 14 ALS-VANCOUVER - Na2O2-fusion ICP-AES

13.5 FLORIN ANALYTICAL - Na2O2-fusion ICP

13 FLORIN ANALYTICAL 2A - 2-Acid ICP GENALYSIS / INTERTEK - Na2O2-fusion ICP-AES 12.5 KALASSAY - INSPECTORATE - Na2O2-fusion ICP-MS 12 ALEX STEWART-MENDOZA - Na2O2-fusion ICP 0 10 20 30 40 50

MEAN 14.50 STD DEV 0.59 COUNT 36

STANDARDS – QAQC SUMMARY TABLES

Summary of Boron Analytical Quality Control Data Produced by the Company for the Project

2011/ 2012 Exploration Program

Balkan Gold Analytical Quality Control Data – 2011/12 Exploration Program Count Total (%) Comment Sampling Program Boron Sample Count 780 Coarse Blanks 59 7.6% Sourced from Shea Clark Smith, Mineral Exploration and CRM Samples 32 4.1% Geochemistry, Reno, USA Field duplicates 14 1.8% Total QC Samples 105 13.5%

2015 Exploration Program

Balkan Gold Analytical Quality Control Data – 2015 Exploration Program Count Total (%) Comment Sampling Program Boron Borate Sample Count 768 Coarse Blanks 39 5.1% Sourced from Shea Clark Smith, Mineral Exploration and CRM Samples 38 4.9% Geochemistry, Reno, USA Field duplicates 38 4.9% Total QC Samples 115 15.0%

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page A5 of A13 SRK Consulting Piskanja MRE 2016 – Technical Appendix A

Summary results of the standard submissions used in the QAQC programme for Boron

2011/ 2012 Exploration Programme (Volumetric Titration)*

Standard Type Count Assigned Mean Variance Maximum Minimum SRK Comment Code Low 1X B SGS LAKEFIELD 6 6.17 5.83 -5.5% 5.88 5.81 SGS LAKEFIELD 1 Anomaly Mid 2X B 7 11.21 10.69 -4.6% 10.75 10.50 A/R Excluded High 3X B SGS LAKEFIELD 8 14.50 13.95 -3.8% 14.01 13.83

*Results include ARD standard submissions that were re-assayed using titration

2011/ 2012 Exploration Programme (Aqua regia digest ICP-MS)

Standard Type Count Assigned Mean Variance Maximum Minimum SRK Comment Code Low 1X B SGS LAKEFIELD 7 6.17 6.51 5.5% 9.91 3.19 Mid 2X B SGS LAKEFIELD 7 11.21 13.58 21.1% 17.52 7.51 SGS LAKEFIELD 1 Anomaly High 3X B 7 14.50 17.25 19.0% 23.32 9.87 A/R Excluded

2015 Exploration Programme (Na2O2-fusion ICP-OES)

Standard Code Lab Count Assigned Mean Variance Maximum Minimum Low 1X B SGS ANKARA 12 6.17 5.74 -7% 5.96 5.33 Mid 2X B SGS ANKARA 13 11.21 10.70 -5% 10.90 10.40 High 3X B SGS ANKARA 13 14.50 14.05 -3% 15.00 13.10

STANDARDS – QAQC CHARTS

2011/ 2012 Exploration Program

Volumetric titration

Analysis B Assays (%) from CRM-Low 1X B Analysis B Assays (%) from CRM-Mid 2X B 8.0 15.0

14.0

13.0 7.0 12.0 +3SD +3SD -3SD 11.0 -3SD +2SD +2SD 10.0

6.0 -2SD -2SD B Assay (%) B Assay (%) Certified Value 9.0 Certified Value SGS LAKEFIELD SGS LAKEFIELD 8.0 ALS ROMAINA ALS ROMAINA 5.0 7.0 SGS LAKEFIELD A/R

6.0

4.0 5.0 0 2 4 6 8 10 0 2 4 6 8 10 Assay Number Assay Number Analysis B Assays (%) from CRM-High 3X B 18.0

17.0

16.0 +3SD 15.0 -3SD +2SD 14.0 -2SD

B Assay (%) Certified Value 13.0 SGS LAKEFIELD ALS ROMAINA 12.0

11.0

10.0 0 2 4 6 8 10 Assay Number

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page A6 of A13 SRK Consulting Piskanja MRE 2016 – Technical Appendix A

Aqua regia digest ICP-MS

Analysis B Assays (%) from CRM-Low 1X B Analysis B Assays (%) from CRM-Mid 2X B 11.0 19.0 18.0 10.0 17.0 16.0 9.0 15.0 8.0 14.0 +3SD 13.0 +3SD 7.0 -3SD 12.0 -3SD 11.0 6.0 +2SD +2SD 10.0 -2SD -2SD

5.0 9.0 B Assay (%) Certified Value B Assay (%) 8.0 Certified Value 4.0 SGS LAKEFIELD 7.0 SGS LAKEFIELD 6.0 ALS ROMAINA ALS ROMAINA 3.0 5.0 4.0 2.0 3.0 1.0 2.0 1.0 0.0 0.0 0 2 4 6 8 10 0 2 4 6 8 10 Assay Number Assay Number Analysis B Assays (%) from CRM-High 3X B 25.0

20.0

+3SD

15.0 -3SD +2SD -2SD

B Assay (%) 10.0 Certified Value SGS LAKEFIELD ALS ROMAINA

5.0 SGS LAKEFIELD A/R

0.0 0 2 4 6 8 10 Assay Number

2015 Exploration Program

Na2O2-fusion ICP-OES

Analysis B Assays (%) from CRM-Low 1X B Analysis B Assays (%) from CRM-Mid 2X B

8.0 14.0

13.0

7.0

+3SD 12.0 +3SD -3SD -3SD +2SD +2SD

6.0 -2SD 11.0 -2SD B Assay BAssay (%) B Assay (%) Certified Value Certified Value SGS ANKARA SGS ANKARA 10.0 BVM PERTH BVM PERTH 5.0

9.0

4.0 8.0 0 5 10 15 0 5 10 15 Assay Number Assay Number Analysis B Assays (%) from CRM-High 3X B 18

17

16 +3SD 15 -3SD +2SD 14 -2SD

B Assay (%) Certified Value 13 SGS ANKARA BVM PERTH 12

11

10 0 5 10 15 Assay Number

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page A7 of A13 SRK Consulting Piskanja MRE 2016 – Technical Appendix A

BLANKS – QAQC CHARTS

2011/ 2012 Exploration Program

Aqua regia digest ICP-MS

Analysis B Assays (%)-BLK

0.500

0.450

0.400

0.350 DL 0.300 3DL 0.250

B Assay BAssay (%) 0.200 5DL

0.150 SGS LAKEFIELD 0.100

0.050

0.000 0 10 20 30 40 50 60 Assay Number

2015 Exploration Program

Na2O2-fusion ICP-OES

Analysis B Assays (%)-BLK

0.300

0.250

0.200 DL

3DL 0.150

B Assay BAssay (%) 5DL 0.100 SGS ANKARA

0.050

0.000 0 10 20 30 40 Assay Number

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page A8 of A13 SRK Consulting Piskanja MRE 2016 – Technical Appendix A

DUPLICATES – QAQC CHARTS

2011/ 2012 Exploration Program

Volumetric titration

Analysis B Assays (%)- Field Duplicates

17

16

15

14

13 y = 0.8515x + 2.3395 R² = 0.9646 12

11

10 X=Y 9 B - SGS_Lakefield

8 BDuplicate (%) 7 Linear (B - SGS_Lakefield)

6

5

4

3

2

1

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 B Original (%)

Aqua regia digest ICP-MS

Analysis B Assays (%)-Original Sample vs Field Duplicates

20

19

18

17

16 y = 1.1372x - 0.2076 15 R² = 0.9722 14

13

12 X=Y 11

10 B - SGS_Lakefield

9 BDuplicate (%) 8 Linear (B - SGS_Lakefield)

7

6

5

4

3

2

1

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 B Original (%)

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page A9 of A13 SRK Consulting Piskanja MRE 2016 – Technical Appendix A

2015 Exploration Program

Na2O2-fusion ICP-OES

Analysis B Assays (%)-Original Sample vs Field Duplicates

15

14

13 y = 0.9932x + 0.0956 12 R² = 0.9871

11

10

9 X=Y

8 B SGS Ankara

7 B - SGS_Ankara (Anomaly Removed)

BDuplicate (%) Linear (B - SGS_Ankara (Anomaly 6 Removed))

5

4

3

2

1

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 B Original (%)

UMPIRE LABORATORY ANALYSIS

2011/ 2012 Exploration Program

Sample Duplicates – SGS Lakefield Titration vs. ALS Romania Na2O2-fusion

Analysis B Assays (%) - Titration vs Fusion- Sample Duplicate

17

16

15

14

13 y = 1.0319x + 0.4083 R² = 0.9834 12

11

10 X=Y

9

8 Original (SGS Lakefield)vsOriginal (ALS BDuplicate (%) 7 Romainia) Linear (Original (SGS 6 Lakefield)vsOriginal (ALS Romainia)) 5

4

3

2

1

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 B Original (%)

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page A10 of A13 SRK Consulting Piskanja MRE 2016 – Technical Appendix A

Sample Duplicates – SGS Lakefield ARD vs. ALS Romania ARD

Analysis B Assays (%) - ARD vs ARD - Sample Duplicate

20

19

18

17

16 y = 0.7756x + 0.1551 15 R² = 0.963 14

13

12 X=Y 11

10 Original (SGS 9 Lakefield)vsDuplicate (ALS BDuplicate (%) Romainia) 8 Linear (Original (SGS 7 Lakefield)vsDuplicate (ALS 6 Romainia))

5

4

3

2

1

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 B Original (%)

Standards – ALS Romania Na2O2-fusion

Analysis B Assays (%) from CRM-Low 1X B Analysis B Assays (%) from CRM-Mid 2X B 8.0 14.0

13.0

7.0 +3SD 12.0 +3SD -3SD -3SD +2SD +2SD 6.0

-2SD 11.0 -2SD B Assay (%) Certified Value B Assay (%) Certified Value SGS LAKEFIELD SGS LAKEFIELD 10.0 ALS ROMAINA ALS ROMAINA 5.0

9.0

4.0 8.0 0 2 4 6 8 10 0 2 4 6 8 10 Assay Number Assay Number Analysis B Assays (%) from CRM-High 3X B 16.5

16.0

15.5

15.0 +3SD -3SD 14.5 +2SD -2SD 14.0

B Assay (%) Certified Value SGS LAKEFIELD 13.5 ALS ROMAINA 13.0

12.5

12.0 0 2 4 6 8 10 Assay Number

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page A11 of A13 SRK Consulting Piskanja MRE 2016 – Technical Appendix A

2015 Exploration Program

Sample Duplicates – SGS Ankara Na2O2-fusion vs. ALS Romania Na2O2-fusion

Analysis B Assays (%)-SGS Ankara vs BVM Perth

20

19

18

17

16

15

14 y = 0.9677x - 0.0209 R² = 0.9965 13

12 X=Y 11 BVM Perth Vs. SGS 10 Ankara

9 BVMPerth B(%) 8

7

6

5

4

3

2

1

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 SGS Ankara B (%)

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page A12 of A13 SRK Consulting Piskanja MRE 2016 – Technical Appendix A

LABORATORY VERIFICATION

2015 Exploration Program

Sample Duplicates – SGS Lakefield Titration vs. ALS Romania Na2O2-fusion

Analysis B Assays (%)-SGS Ankara vs SGS Lakefield

20

19

18

17

16

15 y = 0.9344x - 0.0743 14 R² = 0.9858 13

12 X=Y 11 SGS Lakefield Vs. SGS 10 Ankara 9

SGS SGS Lakefield (%)B 8

7

6

5

4

3

2

1

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 SGS Ankara B (%)

Sample Duplicates – SGS Lakefield Titration vs. BVM Perth Na2O2-fusion

Analysis B Assays (%)-BVM Perth vs SGS Lakefield

20

19

18

17

16

15

14 y = 1.0279x - 0.2932 R² = 0.991 13

12 X=Y 11 SGS Lakefield Vs. BVM 10 Perth 9

SGS SGS Lakefield (%)B 8

7

6

5

4

3

2

1

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 BVM B (%)

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page A13 of A13 SRK Consulting Piskanja MRE 2016 – Technical Appendix B

APPENDIX

B MINERAL RESOURCE ESTIMATE

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page B1 of B9 SRK Consulting Piskanja MRE 2016 – Technical Appendix B

RAW AND LOG HISTOGRAMS

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page B2 of B9 SRK Consulting Piskanja MRE 2016 – Technical Appendix B

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page B3 of B9 SRK Consulting Piskanja MRE 2016 – Technical Appendix B

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page B4 of B9 SRK Consulting Piskanja MRE 2016 – Technical Appendix B

QKNA/ SENSITIVITY ANALYSIS

QKNA Search Ellipse Dimension: Borate horizons domain (KZONE3) for borate

DETERMINE SEARCH VOLUME GRADE RUN Min Max Search SVOL B2O3OK B2O3IDW SLOPE NUM KV % Fill 6 20 50x50x10 1 32.33 32.45 0.64 7 126.22 23.7% 1 6 20 50x50x10 2 28.89 29.56 0.65 12 133.51 46.3% 2 20 50x50x10 3 31.19 30.96 0.50 12 156.12 30.0% 6 20 65x65x10 1 31.09 31.15 0.65 8 126.78 44.0% 2 6 20 65x65x10 2 29.51 29.92 0.58 12 144.25 39.0% 2 20 65x65x10 3 31.93 31.18 0.48 12 156.61 17.0% 6 20 85x85x10 1 30.00 30.48 0.68 11 125.82 58.9% 3 6 20 85x85x10 2 30.99 31.06 0.54 13 151.31 31.6% 2 20 85x85x10 3 30.73 29.53 0.52 15 148.40 9.5% 6 20 105x105x10 1 29.84 30.48 0.69 13 127.36 69.6% 4 6 20 105x105x10 2 31.42 31.23 0.53 12 151.19 26.1% 2 20 105x105x10 3 29.38 28.91 0.44 16 157.36 4.3% 6 20 125x125x10 1 29.86 30.47 0.69 15 128.69 78.4% 5 6 20 125x125x10 2 31.74 31.53 0.50 12 154.36 19.4% 2 20 125x125x10 3 26.72 27.12 0.43 15 157.33 2.2%

QKNA Number of Samples: Borate horizons domain (KZONE3) for borate

DETERMINE NUMBER OF SAMPLES GRADE RUN Min Max MaxKey Search SVOL B2O3OK B2O3IDW SLOPE NUM KV % Fill 8 20 2 65x65x10 1 31.64 31.84 0.69 9 120.473 30.1% 1 8 20 2 65x65x10 2 28.44 29.41 0.68 15 132.146 43.6% 2 20 2 65x65x10 3 32.08 31.23 0.52 14 153.055 26.2% 10 20 2 65x65x10 1 32.46 32.98 0.76 11 106.892 12.6% 2 10 20 2 65x65x10 2 28.71 29.73 0.74 17 125.459 55.5% 2 20 2 65x65x10 3 31.39 30.73 0.54 15 150.846 31.9% 6 30 2 65x65x10 1 31.09 31.15 0.65 8 126.779 44.0% 3 6 30 2 65x65x10 2 29.52 29.92 0.58 12 144.217 39.0% 2 30 2 65x65x10 3 31.87 31.18 0.48 14 156.296 17.0% 6 40 2 65x65x10 1 31.09 31.15 0.65 8 126.779 44.0% 4 6 40 2 65x65x10 2 29.52 29.92 0.58 12 144.217 39.0% 2 40 2 65x65x10 3 31.86 31.16 0.48 14 156.257 17.0% 6 20 0 65x65x10 1 31.94 32.93 0.56 16 146.566 72.4% 5 6 20 0 65x65x10 2 30.86 30.15 0.46 16 161.270 21.4% 2 20 0 65x65x10 3 31.42 31.37 0.40 14 166.26 6.2%

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page B5 of B9 SRK Consulting Piskanja MRE 2016 – Technical Appendix B

VALIDATION PLOTS – BORATE GRADE

KZONE1

50 80 45 40

45 70 40 35

40 35 60 30 35 30 50 30 25 25

25 40 20 B2O3% B2O3% 20

20 No.Samples 30 15 No.Samples 15 15 20 10 10 10 10 5 5 5

0 0 0 0 7471400 7471600 7471800 7472000 7472200 7472400 7472600 4803700 4803800 4803900 4804000 4804100 4804200 4804300 4804400 X-Co-ordinate Y Co-ordinate Sample Mean Model Mean No. Samples Sample Mean Model Mean No. Samples 50 25

45

40 20

35

30 15

25 B2O3%

20 10 No.Samples

15

10 5

5

0 0 -100 -50 0 50 100 150 200 250 300 Z Co-ordinate Sample Mean Model Mean No. Samples

KZONE2

70 25 45 30

40 60 25 20 35 50 30 20 15 40 25

15 B2O3% 30 B2O3% 20

10 No.Samples No.Samples 15 10 20 5 10 10 5 5

0 0 0 0 7471500 7471600 7471700 7471800 7471900 7472000 7472100 7472200 7472300 4803700 4803800 4803900 4804000 4804100 4804200 4804300 X-Co-ordinate Y Co-ordinate Sample Mean Model Mean No. Samples Sample Mean Model Mean No. Samples 70 14

60 12

50 10

40 8

B2O3% 30 6 No.Samples

20 4

10 2

0 0 0 50 100 150 200 250 300 Z Co-ordinate Sample Mean Model Mean No. Samples

KZONE3

50 120 45 60

45 40 100 50 40 35 35 80 30 40 30 25

25 60 30 B2O3%

B2O3% 20 No.Samples 20 No.Samples 40 15 20 15

10 10 20 10 5 5

0 0 0 0 7471500 7471600 7471700 7471800 7471900 7472000 7472100 7472200 7472300 7472400 7472500 4803700 4803800 4803900 4804000 4804100 4804200 4804300 4804400 X-Co-ordinate Y Co-ordinate Sample Mean Model Mean No. Samples Sample Mean Model Mean No. Samples 45 60

40 50 35

30 40

25 30

B2O3% 20 No.Samples 15 20

10 10 5

0 0 150 200 250 300 350 400 Z Co-ordinate Sample Mean Model Mean No. Samples

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page B6 of B9 SRK Consulting Piskanja MRE 2016 – Technical Appendix B

KZONE4

60 25 45 30 40 50 25 20 35

40 30 20 15 25 30 15

B2O3% 20

B2O3% No.Samples

10 No.Samples 20 15 10

10 5 10 5 5

0 0 0 0 7471900 7471950 7472000 7472050 7472100 7472150 7472200 7472250 7472300 7472350 7472400 4803800 4803850 4803900 4803950 4804000 4804050 4804100 4804150 4804200 X-Co-ordinate Y Co-ordinate Sample Mean Model Mean No. Samples Sample Mean Model Mean No. Samples 45 30

40 25 35

30 20

25 15

B2O3% 20 No.Samples 15 10

10 5 5

0 0 200 220 240 260 280 300 320 340 360 380 400 Z Co-ordinate Sample Mean Model Mean No. Samples

KZONE5

45 3.5 40 3.5

40 3 35 3

35 30 2.5 2.5 30 25 25 2 2

20 B2O3%

B2O3% 20 1.5 1.5 No.Samples No.Samples 15 15 1 1 10 10 0.5 0.5 5 5

0 0 0 0 7472150 7472200 7472250 7472300 7472350 7472400 7472450 7472500 4803750 4803800 4803850 4803900 4803950 4804000 4804050 4804100 4804150 4804200 4804250 X-Co-ordinate Y Co-ordinate Sample Mean Model Mean No. Samples Sample Mean Model Mean No. Samples 45 3.5

40 3 35 2.5 30

25 2

B2O3% 20 1.5 No.Samples 15 1 10 0.5 5

0 0 200 220 240 260 280 300 320 340 Z Co-ordinate Sample Mean Model Mean No. Samples

KZONE6

45 35 45 30

40 40 30 25 35 35 25 30 30 20

25 20 25

15 B2O3%

B2O3% 20 15 20

No.Samples No.Samples 15 15 10 10 10 10 5 5 5 5

0 0 0 0 7471600 7471700 7471800 7471900 7472000 7472100 7472200 7472300 7472400 7472500 4803800 4803850 4803900 4803950 4804000 4804050 4804100 4804150 4804200 4804250 4804300 4804350 X-Co-ordinate Y Co-ordinate Sample Mean Model Mean No. Samples Sample Mean Model Mean No. Samples 40 20

18 35 16 30 14 25 12

20 10 B2O3%

8 No.Samples 15 6 10 4 5 2

0 0 0 50 100 150 200 250 300 350 400 Z Co-ordinate Sample Mean Model Mean No. Samples

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page B7 of B9 SRK Consulting Piskanja MRE 2016 – Technical Appendix B

KZONE7

25 14 20 16 18 12 14 20 16 12 10 14 10 15 12 8

10 8 B2O3% B2O3% 6

8 No.Samples 10 No.Samples 6 4 6 4 5 4 2 2 2

0 0 0 0 7471650 7471700 7471750 7471800 7471850 7471900 4803800 4803850 4803900 4803950 4804000 4804050 X-Co-ordinate Y Co-ordinate Sample Mean Model Mean No. Samples Sample Mean Model Mean No. Samples 25 14

12 20 10

15 8

B2O3% 6

10 No.Samples

4 5 2

0 0 0 20 40 60 80 100 120 140 160 180 Z Co-ordinate Sample Mean Model Mean No. Samples

KZONE8

35 4.5 35 6

4 30 30 5 3.5 25 25 3 4

20 2.5 20

3 B2O3%

B2O3% 15 2 15

No.Samples No.Samples 1.5 2 10 10 1 1 5 5 0.5

0 0 0 0 7472200 7472250 7472300 7472350 7472400 7472450 4803750 4803800 4803850 4803900 4803950 4804000 4804050 X-Co-ordinate Y Co-ordinate Sample Mean Model Mean No. Samples Sample Mean Model Mean No. Samples 35 3.5

30 3

25 2.5

20 2

B2O3% 15 1.5 No.Samples

10 1

5 0.5

0 0 180 200 220 240 260 280 300 Z Co-ordinate Sample Mean Model Mean No. Samples

KZONE9

60 3.5 60 6

50 3 50 5

2.5 40 40 4 2

30 30 3 B2O3%

1.5 B2O3%

No.Samples No.Samples 20 20 2 1

10 0.5 10 1

0 0 0 0 7471500 7471550 7471600 7471650 7471700 7471750 7471800 7471850 7471900 4803900 4803950 4804000 4804050 4804100 4804150 X-Co-ordinate Y Co-ordinate Sample Mean Model Mean No. Samples Sample Mean Model Mean No. Samples 60 3.5

50 3

2.5 40 2 30

B2O3% 1.5 No.Samples 20 1

10 0.5

0 0 -80 -60 -40 -20 0 20 40 60 80 100 120 140 Z Co-ordinate Sample Mean Model Mean No. Samples

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page B8 of B9 SRK Consulting Piskanja MRE 2016 – Technical Appendix B

KZONE10

30 4.5 18 6

4 16 25 5 3.5 14

20 3 12 4

2.5 10

15 3 B2O3%

2 B2O3% 8

No.Samples No.Samples 10 1.5 6 2

1 4 5 1 0.5 2

0 0 0 0 7471600 7471620 7471640 7471660 7471680 7471700 7471720 7471740 7471760 7471780 7471800 4803750 4803800 4803850 4803900 4803950 4804000 X-Co-ordinate Y Co-ordinate Sample Mean Model Mean No. Samples Sample Mean Model Mean No. Samples 30 4.5

4 25 3.5

20 3

2.5 15

B2O3% 2 No.Samples 10 1.5

1 5 0.5

0 0 180 190 200 210 220 230 240 250 260 270 280 Z Co-ordinate Sample Mean Model Mean No. Samples

U6467 Piskanja MRE 2016_Final Report.docx November, 2016 Page B9 of B9