TSXV: CSG
OTCQB: COSRF
Technical Brief · Taraira North (Machado)

Coarse Gold at Taraira: A Brief for Geologists

Written for geologists and technical teams evaluating Cosigo Resources' Taraira North (Machado) gold project in the Vaupés region of Colombia: the project's key value drivers, the caveats that belong alongside them, and the data behind both.

The Cosigo Taraira gold system shows a coarse, nuggety gold distribution accompanied by a Te-Bi(-Ag-W-Cu) geochemical signature rather than the classic As-Sb pathfinder suite, pointing toward a telluride-bearing or intrusion-related gold style rather than disseminated sulfide-hosted mineralization.

This is supported by three independent lines of evidence from the 2013 RC program: ALS metallic screen fire assay (Au-SCR24) shows the coarse (+) fraction — typically only 2–3% of sample weight — carrying a median of 6% and up to 83% of total gold mass in gold-bearing intervals; tellurium is by far the strongest correlate of gold occurrence (Spearman r ≈ 0.51, 8× median enrichment when gold is present); and, using the one method/lab common across both drilling types (ALS Au-AA26), RC assays run systematically 1.3–2× higher than core across every percentile of the grade distribution, despite no lab or method difference.

Taken together, this suggests a real nugget effect intrinsic to the deposit — likely exacerbated in RC by surface crushing liberating and homogenizing coarse gold into the sample pulp more effectively than core cutting does — with the practical implication that conventional small-aliquot fire assay grades, particularly from RC, should be treated as a probable floor rather than a reliable point estimate, and that metallic screen assay or bulk/metallurgical-scale testing will be needed for confident grade estimation as the project advances.

83%
Max share of gold in the coarse screen fraction (2–3% of mass)
r ≈ 0.51
Te–Au Spearman correlation, 2013 RC
1.3–2×
RC vs core, same ALS Au-AA26 method
~88%
Gravity recovery, July 2010 Knelson test
Summary

Key Value Drivers at a Glance

Gold character
Gold at Cosigo Resources' Taraira North (Machado) project is coarse and nuggety. In the 2013 RC program the coarse (+) screen fraction — typically only 2–3% of sample weight — carried a median of 6% and up to 83% of the gold in gold-bearing intervals (ALS Au-SCR24).
Geochemical signature
Gold at Taraira is accompanied by a Te-Bi(-Ag-W-Cu) signature rather than the classic As-Sb pathfinder suite. Tellurium is the strongest correlate of gold in the 2013 RC data (Spearman r ≈ 0.51, 8× median enrichment when gold is present).
Screen-fire vs fire assay
On the 748 Taraira RC samples assayed both ways, screen-fire (Au-SCR24) totals ran about 22–25% higher than routine fire assay (Au-AA26), while the median sample differed by only about 3% — the signature of a nugget effect.
RC vs core
On the same ALS Au-AA26 method, Taraira RC assays run 1.3–2× higher than core assays across every percentile of the grade distribution.
Gravity recovery
Three Knelson gravity tests have been run on Taraira material. The July 2010 test on high-grade tunnel rock recovered 91.5% and 84.4% of the gold by gravity (about 88% on average); the March 2010 test recovered 67.9–70.7%; and a 2026 test on low-grade conglomerate put 64.9% of the gold in the Knelson concentrate.
2010 channel vs bulk sample
Two channel samples chipped from a face at Agamenon's adit just before it was blasted assayed 0.73 and 0.98 g/t Au; the bulk sample from that blast returned calculated head grades of 37.6 and 27.2 g/t Au — roughly 30 to 50 times higher.
Practical implication
Conventional small-aliquot fire assay grades from Taraira, particularly from RC, should be treated as a probable floor rather than a reliable point estimate; metallic screen assay or bulk/metallurgical-scale testing will be needed for confident grade estimation.
Resource status
Exploration stage. Cosigo Resources has not defined a mineral resource or mineral reserve under NI 43-101.
Metallurgy

Gold is highly amenable to gravity recovery — no expensive chemistry required

The July 2010 Knelson bulk sample test (KRTS 20564) is the single most important metallurgical data point for this project so far.

What was tested

A bulk sample from a blasted underground round at Agamenon's adit in the central Machado (Cerro Rojo) workings — 617 kg from two randomly selected barrels of an 18-barrel bulk sample — was crushed, blended, and split into two sub-samples of about 10 kg. Each was ground to P80 ~100 µm and run through a laboratory Knelson KC-MD3 concentrator at 60 G with 3.5 L/min of fluidization water. The Knelson concentrate was panned; the pan concentrate and pan tails were fire assayed to extinction, and the tails were sub-sampled and assayed in duplicate, giving a full metallurgical balance for each sub-sample. The method is specifically designed to avoid the sub-sampling "nugget effect" that distorts conventional small-sample assays.

Source

Knelson Research & Technology Centre, report "KRTS 20564 — Cosigo Resources Gravity Grade Test Report," prepared by Mike Tran for Dennis Milburn, July 9, 2010 (company records). Publicly reported in a Horseshoe Gold Mining Inc. news release dated August 18, 2010, reviewed by Joseph H. Montgomery, PhD, P.Eng., a Qualified Person under NI 43-101. The 2011 NI 43-101 technical report includes the results as Appendix III.

KRTS 20564 results (July 9, 2010). Grind P80 ~100 µm. Recovery and head grade are back-calculated from the metallurgical balance of each sub-sample.
Sub-sampleMass (kg)Calculated head grade (g/t Au)Gravity recovery (%)
A10.1637.691.5
B10.5127.284.4
Average~88

Why this matters

Roughly 88% of the gold in this bulk sample could be recovered using gravity concentration alone — no cyanide leaching or other costly chemical processing. This is a major economic/metallurgical advantage: simpler, cheaper processing flowsheet, lower capex/opex, faster path to production, and reduced environmental footprint compared to a project requiring chemical extraction.

Caveat to mention alongside it

Of the gold not recovered by gravity, 64.1% of that loss was concentrated in the finest fraction (below 25 µm) — consistent with a practical recovery ceiling at that grind size rather than an easily-fixed process inefficiency. Finer grinding or a secondary recovery step (e.g. flotation of the gravity tails) could potentially push recovery higher, but that has not been tested.

Every gravity test on Taraira material

The July 2010 result comes from one location, Agamenon's adit at Cerro Rojo. Two other Knelson tests have been run on Taraira material: a March 2010 test on vein rock from the Chile Bajo adit, also at Cerro Rojo, and a 2026 test on low-grade Chicken Coop (Gallinero) conglomerate. Gravity recovery has not yet been tested at other target areas or drill-hole zones — a "confirm at other targets" item for further work programs.

Knelson gravity tests on Taraira material, 2010–2026.
TestDateMaterialHead gradeGravity recovery
KRTS 20542March 29, 2010Vein rock, Chile Bajo adit, Cerro Rojo (840 kg)6.47–7.11 g/t Au67.9–70.7%
KRTS 20564July 9, 2010Tunnel rock, Agamenon's adit, Cerro Rojo (617 kg)27.2–37.6 g/t Au84.4–91.5% (avg ~88%)
Chicken Coop conglomerateReported April 7, 2026Conglomerate composite, Chicken Coop (Gallinero) unit (60 kg; one-pass 20 kg Knelson test)0.03 g/t Au64.9% of the gold in the Knelson concentrate*

* The 2026 release reports the pan concentrate holding 6.9% and the pan tails 58.0% of the gold in the 20 kg sample — together, the Knelson concentrate. At a head grade of 0.03 g/t Au, that result says little about high-grade material.

Source: KRTS 20564 Cosigo Resources — Gravity Grade Test Report · Horseshoe Gold Mining Inc. news release reporting the Knelson gravity test results · Technical Report, Machado Project, Vaupés Department, Colombia · KRTS 20542 Cosigo Resources — Gravity Grade Test Report · Cosigo Resources Announces Results of Conglomerate Layer Bulk Sample for Gravity Concentration Tests

Sampling · 2010

Origin of the Nugget-Effect Awareness: Channel Samples vs Bulk Sample

The same 2010 program is also the project's first clear evidence of a strong nugget effect, which now shapes how all sampling and assay data at Taraira should be interpreted.

Two conventional channel samples were chipped from the same underground face right before the blast that produced the July 2010 bulk sample. Their assays came back very low grade: 0.73 ppm (g/t) Au for the #1 channel sample and 0.98 ppm for the #2 channel sample, assayed in Vancouver. The bulk sample taken from that same blast — processed as a full metallurgical sample rather than a small chip — returned calculated head grades of 37.6 g/t and 27.2 g/t Au, roughly 30 to 50 times higher (about 38 times comparing the averages).

This gap is what first alerted Andy Rendle, now Cosigo's President & COO, and the Company to the likelihood of a strong nugget effect at this project: small conventional samples (channel chips, standard assay aliquots) can badly understate true grade when coarse, erratically-distributed gold particles are present, because a small sample has a high chance of simply missing the particles. The 2011 NI 43-101 technical report drew the same conclusion from the bulk samples, describing Cerro Rojo as "a coarse gold (nuggety) mineral occurrence." The pattern was later reinforced by the 2013 RC program's routine fire assay vs screen-fire assay comparison.

Same face, same blast: conventional channel samples vs the processed bulk sample.
SampleTypeAu (g/t)
#1 channel sampleConventional channel chip, pre-blast face0.73
#2 channel sampleConventional channel chip, pre-blast face0.98
KRTS 20564-ABulk sample from the same blast — calculated head grade37.60
KRTS 20564-BBulk sample from the same blast — calculated head grade27.20

Practical implication

Grades from small conventional samples (channel chips, standard 30–50 g assay aliquots) at this project should be treated as a likely floor, not a reliable estimate, especially in higher-grade zones. Bulk sampling, screen-fire assay, or metallurgical-scale testing are more trustworthy for true grade estimation. This should be flagged whenever conventional assay results are discussed or compared across the project.

Source: 2010 pre-blast channel samples, Agamenon's adit face, Cerro Rojo · KRTS 20564 Cosigo Resources — Gravity Grade Test Report · Technical Report, Machado Project, Vaupés Department, Colombia

Assay Methods · 2013 RC

Routine Fire Assay (Au‑AA26) vs Screen-Fire Assay (Au‑SCR24)

Computed directly from Cosigo's Taraira North master sample database (all 1,032 rows) on September 2, 2026, to put a hard number on the pattern described above.

Sample coverage

748 samples have both a routine fire assay (ALS Au-AA26) and a screen-fire assay (ALS Au-SCR24 total, coarse + fine fractions recombined) on file — all from the 2013 RC drilling program (holes 13MR001–13MR020); no other campaign ran both methods on the same samples. Of those, 265 samples returned a real detectable value on both methods (neither result reported as below the lab's detection limit). That 265-sample set is the cleanest basis for magnitude comparisons.

Where the gold sits

The metallic screen assay also shows where the gold sits. In gold-bearing intervals, the coarse (+) fraction — typically only 2–3% of sample weight — carried a median of 6% and up to 83% of the total gold.

Direction — which method reads higher

Share of paired samples in which each method returned the higher gold value.
Sample setScreen-fire higherFire assay higherTied
All 748 pairs (below-detection results at half the detection limit)71.8%20.7%7.5%
265 pairs with a real value on both methods51.3%27.5%21.1%

Screen-fire reads higher far more often than lower, but not always. Even on the strictest comparison, screen-fire beats routine fire assay roughly 2-to-1 when they differ — but fire assay comes out higher in a meaningful minority of samples, so it isn't a one-way rule at the individual-sample level.

Magnitude — in aggregate, screen-fire runs about 22–25% higher

Sum of screen-fire results divided by sum of routine fire assay results.
Sample setΣ SCR24 / Σ AA26Screen-fire higher by
265 pairs with a real value on both methods1.2222%
747 pairs (fuller set)1.2525%

In aggregate, screen-fire runs about 22–25% higher — both figures consistent with the ~20% figure previously noted in Cosigo's review of the 2013 screen-fire assays. The average is pulled up by a handful of samples with a large gap, which is itself the nugget-effect signature: most samples are close between the two methods (median ratio ~1.03, only about 3% apart), but a minority run dramatically higher by screen-fire. At the low-grade end, several samples show screen-fire reading 5–8× higher than fire assay in relative terms.

Largest individual gaps

Individual 2013 RC samples where screen-fire and routine fire assay diverge most. Each sample is a 1.52 m (5 ft) RC interval.
HoleSampleAu-SCR24 (g/t)Au-AA26 (g/t)Difference (g/t)RatioNote
13MR007P94048829.9018.2511.651.6×Largest gap in the dataset
13MR011P9402078.322.395.933.5×Next largest; the screen-fire value is the reported 1.52 m @ 8.32 g/t Au sub-interval
13MR020P9407480.170.020.158.5×Low-grade example of a large relative gap

Fire assay (AA26) was not uniformly lower — but it usually was, and when it was lower it was sometimes dramatically lower on individual samples, while running ~20–25% low in aggregate across the whole 2013 RC dataset. That combination (median close, aggregate skewed by a few big outliers) is exactly what a nugget effect looks like, and matches the bulk-sample-vs-channel-sample story from 2010.

Bottom line

One small check pointed the other way. In April 2024 Cosigo reported that a two-sample screen test on core from the Gallinero 1 and 2 test holes "showed no benefit compared to values obtained from standard samples." Those holes were low grade — the release reports gold values of 10 to 380 ppb — where a coarse-gold effect is hard to detect.

Source: Taraira North master sample database — Au-AA26 vs Au-SCR24, coarse-fraction, multi-element and RC-vs-core analysis · Cosigo Drills High Grade Gold-Bearing Strata at Machado · Cosigo Drill Conversion Test and Results

Drilling Methods

RC vs Core on the Same Assay Method

Only one gold method and laboratory is common to both drilling types in Cosigo's database: ALS fire assay with an AA finish, Au-AA26. On that like-for-like basis, Taraira RC assays run systematically 1.3–2× higher than core across every percentile of the grade distribution, despite no lab or method difference.

Cosigo's interpretation is that this reflects a real nugget effect intrinsic to the deposit, likely exacerbated in RC by surface crushing liberating and homogenizing coarse gold into the sample pulp more effectively than core cutting does.

Limitation

The comparison is between different holes, not twinned holes. The RC samples come mainly from the 2013 program (holes 13MR001–13MR020) and the core from Gallinero 1 and 2 (2023), 24TAR_DH_022 (2024) and hole 25, drilled in different years and at different locations. Part of the difference could therefore be spatial; twinning selected RC holes with core would test it directly.

Source: Taraira North master sample database — Au-AA26 vs Au-SCR24, coarse-fraction, multi-element and RC-vs-core analysis · Cosigo Core Results During Hydrogeological Advancement Drilling · Cosigo Resources Provides Drilling Update

Geochemistry

A Te-Bi(-Ag-W-Cu) Signature, Not As-Sb

Gold at Taraira is accompanied by a tellurium–bismuth signature, with silver, tungsten and copper as lesser associates — Te-Bi(-Ag-W-Cu) — rather than the classic arsenic–antimony (As-Sb) pathfinder suite. In the 2013 RC multi-element data, tellurium is by far the strongest correlate of gold occurrence (Spearman r ≈ 0.51), with an 8× median enrichment in tellurium when gold is present.

A working interpretation

Cosigo reads this signature as pointing toward a telluride-bearing or intrusion-related gold style rather than disseminated sulfide-hosted mineralization. It is a working interpretation of the 2013 RC geochemistry. Cosigo's NI 43-101 technical reports describe the Taraira district as having potential for paleoplacer or modified paleoplacer gold (the Witwatersrand-type model) and/or epigenetic vein- and shear-related gold. A telluride-bearing or intrusion-related style would fall within the epigenetic family; the reports' models are unchanged, and Cosigo continues to evaluate the district's conglomerate and stratabound horizons.

Source: Taraira North master sample database — Au-AA26 vs Au-SCR24, coarse-fraction, multi-element and RC-vs-core analysis · Technical Report, Machado Project, Vaupés Department, Colombia

Evaluating the Project

Implications for Grade Estimation

  1. Treat conventional small-aliquot fire assay grades (30–50 g), particularly from RC, as a probable floor rather than a reliable point estimate — especially in higher-grade zones.
  2. Use metallic screen (screen-fire) assay in mineralized intervals: it assays the coarse fraction to extinction instead of relying on one small aliquot.
  3. Use bulk or metallurgical-scale testing for confident grade estimation as the project advances. A Knelson gravity grade test processes each 10–14 kg sub-sample in its entirety.
  4. Compare RC and core grades only with the RC-vs-core bias in mind, and twin selected RC holes with core to separate sampling bias from spatial variation.
  5. Confirm gravity recovery at other target areas and drill-hole zones; the high-grade result so far comes from one location.
  6. Flag the nugget effect whenever conventional assay results are discussed or compared across the project.

Data and limitations

  • The assay statistics on this page were computed by Cosigo from its Taraira North master sample database (25 holes, 1,032 rows) on September 2, 2026. They are a compilation of previously collected laboratory data, not new sampling.
  • In the 748-sample comparison, below-detection results are treated conservatively at half the detection limit; the 265-sample set excludes them.
  • The RC-versus-core comparison is between different holes, not twins, so part of the difference may be spatial.
  • The gravity tests are laboratory-scale tests on material from specific locations and are not representative of the project as a whole.
  • Cosigo Resources is an exploration-stage company and has not defined a mineral resource or mineral reserve under NI 43-101.
Questions

Technical FAQ

Is there a nugget effect at Cosigo Resources' Taraira gold project?

Yes. Gold at Cosigo Resources' Taraira North (Machado) project in Vaupés, Colombia, is coarse and nuggety, and several lines of evidence show a nugget effect. In 2010 two channel samples assayed 0.73 and 0.98 g/t Au, while the bulk sample blasted from the same face returned calculated head grades of 27.2 and 37.6 g/t Au. On 2013 RC samples assayed both ways, screen-fire assays totalled about 22–25% more gold than routine fire assays, and in gold-bearing intervals the coarse screen fraction — only 2–3% of sample weight — carried up to 83% of the gold. Conventional small-aliquot fire assay grades at Taraira are therefore treated as a probable floor rather than a reliable estimate.

Can gold from Cosigo Resources' Taraira project be recovered by gravity?

Laboratory tests show gold from Cosigo Resources' Taraira North (Machado) project is highly amenable to gravity concentration. In the July 2010 Knelson test on 617 kg of tunnel rock from Agamenon's adit at Cerro Rojo, gravity alone recovered 91.5% and 84.4% of the gold from two sub-samples (about 88% on average), at calculated head grades of 37.6 and 27.2 g/t Au. A March 2010 test on vein rock from the Chile Bajo adit recovered 67.9–70.7%, and a 2026 test on low-grade Chicken Coop conglomerate put 64.9% of the gold in the Knelson concentrate. These are laboratory-scale exploration results, and gravity recovery has not yet been tested at other target areas.

How much of the gold at Cosigo Resources' Taraira project is coarse?

In ALS metallic screen fire assays (Au-SCR24) from Cosigo Resources' 2013 reverse-circulation program at Taraira North, the coarse (+) fraction was typically only 2–3% of sample weight but carried a median of 6%, and up to 83%, of the total gold mass in gold-bearing intervals.

How do screen-fire and routine fire assays compare at Cosigo Resources' Taraira project?

Across 748 samples from Cosigo Resources' 2013 Taraira RC program that have both a routine fire assay (ALS Au-AA26) and a screen-fire assay (ALS Au-SCR24), screen-fire read higher in 71.8% of samples, lower in 20.7% and tied in 7.5%. In aggregate screen-fire ran about 22% higher on the 265 pairs with a real value on both methods and about 25% higher on the fuller set, while the median pair differed by only about 3%. The largest single gap was sample P940488 in hole 13MR007: 29.90 g/t Au by screen-fire versus 18.25 g/t Au by fire assay.

Why do RC assays run higher than core assays at Cosigo Resources' Taraira project?

Using the one method and laboratory common to both drilling types (ALS Au-AA26), Cosigo Resources' Taraira RC assays run 1.3–2× higher than core across every percentile of the grade distribution. Cosigo interprets this as a real nugget effect, likely exacerbated in RC because surface crushing liberates and homogenizes coarse gold into the sample pulp more effectively than core cutting does. The RC and core samples come from different holes rather than twinned holes, so part of the difference may be spatial.

What pathfinder elements are associated with gold at Cosigo Resources' Taraira project?

Gold at Cosigo Resources' Taraira North (Machado) project is accompanied by a tellurium–bismuth signature with lesser silver, tungsten and copper — Te-Bi(-Ag-W-Cu) — rather than the classic arsenic–antimony (As-Sb) pathfinder suite. In the 2013 RC multi-element data tellurium is by far the strongest correlate of gold occurrence (Spearman r ≈ 0.51), with an 8× median enrichment in tellurium when gold is present.

What style of gold mineralization does the Taraira geochemistry suggest?

Cosigo Resources reads the Te-Bi(-Ag-W-Cu) signature and coarse gold at Taraira as pointing toward a telluride-bearing or intrusion-related gold style rather than disseminated sulfide-hosted mineralization. This is a working interpretation of the 2013 RC geochemistry. Cosigo's NI 43-101 technical reports describe potential for paleoplacer (Witwatersrand-type) and/or epigenetic vein- and shear-related gold; an intrusion-related style falls within the epigenetic family, and no mineral resource has been defined.

Are conventional fire assay grades from Cosigo Resources' Taraira project reliable?

Because gold at Cosigo Resources' Taraira project is coarse and nuggety, conventional small-aliquot fire assay grades (30–50 g), particularly from RC drilling, should be treated as a probable floor rather than a reliable point estimate. Metallic screen assay or bulk and metallurgical-scale testing will be needed for confident grade estimation as the project advances.

References

Sources

The full exploration record, both NI 43-101 technical reports and the complete 2013 drill results.

Historical work2011 NI 43-101 report (PDF)2024 NI 43-101 report (PDF)Contact the Company
Forward-Looking Statements: This website contains forward-looking statements subject to risks and uncertainties. Actual results may differ materially from those anticipated. Scientific and technical disclosure has been reviewed by a Qualified Person (P.Geo. Erik Ostensoe) in accordance with National Instrument 43-101. The TSX Venture Exchange does not accept responsibility for the adequacy or accuracy of this release. Grab samples are selective and may not be representative of average mineralization.