Ejochi botanical validation: reconcile hormone and reproductive-safety evidence

Produce a source-auditable evidence brief for Ejochi's proposed Sphenocentrum jollyanum / Eurycoma longifolia blend. Determine what three published studies establish about hormones and reproductive outcomes, and which claims cannot be transferred to the proposed blend.

This is a bounded critical extraction, not a systematic review or a request to recommend treatment. Assess these four fixed questions: (C1) Does increased testosterone establish improved reproductive function? (C2) What reproductive harm and recovery signals are actually measured for S. jollyanum? (C3) How directly does the E. longifolia stress study support the project's chronic hormonal rationale? (C4) Do these studies establish synergy, mitigation of S. jollyanum toxicity, or safety of the 5:1 combination?

The project context is frozen here: its July 21 and August 5, 2026 discussions propose complementary acute and chronic botanical effects, identify a reproductive-toxicity concern, and describe combination synergy and toxicity mitigation as unproven. It proposes a 5:1 blend and HPLC reproducibility work. These are research claims, not accepted facts. The four questions above govern this task even if the project page changes.

Submit both required files as actual files in the Submission; external links do not substitute for them.

• evidence.csv: at least two distinct result rows per required paper, including a hormonal and a sperm/reproductive result for each rat paper and cortisol and testosterone results for the human paper. Columns: row_id, source_id, species_population, sample_size_and_unit, extract_identity, exposure_duration, comparator, endpoint, result_and_units, uncertainty_or_significance, source_locator, relevance_to_claim. Historical study exposures may be recorded descriptively; do not turn them into recommendations. Use not reported where a field is absent; do not estimate graph values.
• brief.md: 900–1400 words excluding references and tables. Include one conclusion for each C1–C4, links to supporting CSV row IDs, a comparison of extract/preparation and population differences, and three ranked evidence gaps that prevent stronger conclusions. Each gap must identify the missing evidence and the specific claim it would resolve. Include full citations, URLs, article dates, access dates, and limitations/conflicts or discrepancies found in the papers. Distinguish paper authors' interpretations from measured results and your inference.

Coverage is the three papers below, treated as three studies, not one pooled sample. Extract only each paper's original results, not results cited in its introduction. No additional studies are required or rewarded. A source's internal disagreement must be identified with both locations rather than silently resolved.

The required corpus is the following identified published articles, in their PMC full-text versions available on September 22, 2026; later article revisions and newly published evidence are outside this task. PMC hosts the public full texts for the judging window. Article identity and publication year govern; report any discrepancy between abstract and main text.

• S1: Raji et al. (2006), *Comprehensive assessment of the effect of Sphenocentrum jollyanum root extract on male reproductive activity in albino rats*: https://pmc.ncbi.nlm.nih.gov/articles/PMC5906826/
• S2: *Assessment of Reproductive Toxicity of Hydroethanolic Root Extracts of Caesalpinia benthamiana, Sphenocentrum jollyanum, and Paullinia pinnata* (2021): https://pmc.ncbi.nlm.nih.gov/articles/PMC7941055/ . Only its S. jollyanum arms and their controls are in scope.
• S3: Talbott et al. (2013), *Effect of Tongkat Ali on stress hormones and psychological mood state in moderately stressed subjects*: https://pmc.ncbi.nlm.nih.gov/articles/PMC3669033/

Background project: https://openlabs-git-codex-openlabs-elgora-adapter-bio-xyz.vercel.app/projects/e3d3177a-7a0f-43e7-8699-fd42936c6be4 . Background discussions: https://openlabs-git-codex-openlabs-elgora-adapter-bio-xyz.vercel.app/post/cdc67fff-3203-411f-8eed-f65eb96e61c2 and https://openlabs-git-codex-openlabs-elgora-adapter-bio-xyz.vercel.app/post/fabf6973-9564-4e0a-af2a-0c59c3f715a3 . The embedded context defines coverage; unavailable background pages do not remove required scientific evidence.

Only submissions meeting every acceptance criterion qualify. Compare eligible entries in this priority order: (1) fidelity and precision of source-to-claim links, especially contradictory findings; (2) completeness of the explanation of why the evidence does or does not transfer to the blend; (3) usefulness of the three gaps, meaning each resolves a distinct decision uncertainty without assuming a desired result. Prefer the entry stronger on the first priority where entries materially differ. If still tied, the earlier on-chain Submission wins; equal timestamps are resolved by lower Submission ID. If only one qualifies it wins; if none qualifies the outcome is no_valid_submission.

New animal or human work, clinical advice, dosing recommendations, blend preparation, HPLC replication, invented data, and safety certification are outside scope.