Heavy Metal Index
Ask the IndexAboutIndex Pro

Codex bulk ingest report - 2026-07-27

Scope Requested mode: fast serial source-page ingest, no synthesis / standards / pooling / design.

Author Karen Pendergrass · ORCID 0000-0002-2348-7259Cite this pageVersion historyDownload PDF

Scope

  • Requested mode: fast serial source-page ingest, no synthesis / standards / pooling / design.
  • Roots order: mercury deep dive, ingredient sweep, product sweep, Manual Fetch Kimi, seasonal/geographic auto-fetch, Manual Resourcing, Manual Fetch Discovery.
  • Started root: raw/Research Pull Mercury Deep Dive (/Volumes/HMI-Library/research-pulls/metal-deep-dive-mercury-hg).

Result

  • New source pages committed in this continuation: 2.
  • Commits made:
    • 6bf226424 (ingest: blechinger2026-methylmercury-neurodevelopment-trvs [fresh] from Research Pull Mercury Deep Dive).
    • 19697b379 (ingest: kim2019-jeju-shark-metals [fresh] from Research Pull Mercury Deep Dive).
  • Source pages created:
    • wiki/sources/blechinger2026-methylmercury-neurodevelopment-trvs.md.
    • wiki/sources/kim2019-jeju-shark-metals.md.
  • Routing audit for the Blechinger source exited 0: 4272 product-source rows generated overall, 13 unresolved targets, 1021 malformed source rows, 0 blocking. The new Blechinger row in routing_malformed.csv is advisory only for intentionally empty products / ingredients.
  • Routing audit for the Kim source exited 0: 4273 product-source rows generated overall, 13 unresolved targets, 1021 malformed source rows, 0 blocking. The Kim source added 1 fish-marine-predatory product-route row and did not appear in routing_unresolved.csv or routing_malformed.csv.
  • Stop reason: paper-internal data-integrity issue in the next new candidate PDF.

Previous Stopper Already Surfaced

Candidate PDF:

raw/Research Pull Mercury Deep Dive/Fish marine non-predatory sardines anchovies salmon cod/JECFA methylmercury/Karjalainen__EstimatedIntake_FinalDraft_2012.pdf

SHA-256:

f5b61d4bb4d58f842211949df2ae90d2161dd95e7b83cb5a9c9466257137160b

Prior finding: the artifact appears incomplete/corrupt. It starts at 1. Introduction, lacks a title page, author list, DOI, and publication venue, and rendered pages around Table 1, Table 2, Figure 1, Figure 2, and Figure 3 show captions without the required table/figure bodies. No source page was written for this artifact.

Current Stopper

Candidate PDF:

raw/Research Pull Mercury Deep Dive/Fish marine non-predatory sardines anchovies salmon cod/Codex mercury fish/10-1007-s12011-021-03064-8.pdf

Identity:

  • DOI: 10.1007/s12011-021-03064-8.
  • SHA-256: 521602b29d1bf3340fc3327555a19ffc410ac6a35ab2d07b6a782fba8bfdb166.
  • Title: Metals in Wild and Cultured Dicentrarchus labrax (Linnaeus, 1758) from Fish Markets in Sinop: Consumer's Health Risk Assessment.
  • Authors: Levent Bat, Fatih Sahin, Md. Simul Bhuyan, Elif Arici, Aysah Oztekin.

Identity / duplicate checks performed:

  • DOI grep for 10.1007/s12011-021-03064-8 and DOI-slug grep for 10-1007-s12011-021-03064-8 found no existing wiki/sources/ page.
  • Raw-handle grep for RPHG_10-1007-s12011-021-03064-8 found no existing source page.
  • Checksum grep for 521602b29d1bf3340fc3327555a19ffc410ac6a35ab2d07b6a782fba8bfdb166 found no existing source page.
  • Proposed cite-key greps for bat2022-metals-seabass-sinop and bat2022-metals-dicentrarchus-labrax-sinop found no existing source page.

Data-integrity issue:

  • Table 1 reports cultured European seabass Cu as 0.27 mg/kg wet wt.; Figure 1 monthly cultured Cu bars are also visually centered around this magnitude.
  • The methods give infant fish intake as 13 g/day, infant body weight as 10 kg, and the EDI formula as concentration times fish intake divided by body weight.
  • Using the paper’s own Table 1 concentration, the cultured-Cu infant EDI should be approximately 3.51E-04 mg/kg body wt./day (0.27 mg/kg * 0.013 kg/day / 10 kg).
  • Table 2 instead reports cultured-Cu infant EDI as 3.079E-03, and the corresponding EWI as 2.2E-02. The expected EWI from the paper’s concentration is about 2.46E-03.
  • Table 2’s cultured-Cu children and adults rows, 2.452E-04 and 1.596E-04, are consistent with the 0.27 mg/kg wet wt. concentration. The inconsistency appears localized to the cultured-Cu infant EDI/EWI row and downstream risk calculations.
  • Table 3 reports cultured-Cu infant THQ as 7.70E-02; using the paper’s 4E-02 Cu reference dose and the concentration-consistent infant EDI gives about 8.8E-03. The reported cultured-infant HI (8.45E-01) therefore appears to include the elevated Cu infant THQ.

Disposition:

  • I stopped before writing a wiki/sources/ page for this Bat 2022 candidate, before appending it to data/evidence/audit-queue.csv, before running a source-route audit for it, and before committing any Bat 2022 ingest artifact.
  • Suggested queue disposition: send the Bat 2022 PDF to Claude/human audit for whether to ingest exact source-reported values with an explicit arithmetic-warning note, exclude the affected risk-calculation cells, or replace with a corrected article version if one exists.

Current Stopper - HJBC Canned Fish Mercury Paper

Candidate PDF:

raw/Research Pull Mercury Deep Dive/Fish marine non-predatory sardines anchovies salmon cod/Codex mercury fish/10-15671-hjbc-778023.pdf

Identity:

  • DOI: 10.15671/hjbc.778023.
  • SHA-256: 2b9e932e3aeb1bbb353cd2f90f6cde4d26ab3c2144b7016984a2a28f85db15f2.
  • Title: Validation of Determination by Icp-Oes Method of Mercury Residual Levels in Meat of Canned Fish Put up for Sale in Turkey.
  • Authors: Fatma Hepsag and Tefide Kizildeniz.

Identity / duplicate checks performed:

  • DOI grep for 10.15671/hjbc.778023 and DOI-slug grep for 10-15671-hjbc-778023 found no existing wiki/sources/ page.
  • Raw-handle grep for RPHG_10-15671-hjbc-778023 found no existing source page.
  • Checksum grep for 2b9e932e3aeb1bbb353cd2f90f6cde4d26ab3c2144b7016984a2a28f85db15f2 found no existing source page.
  • The PDF has 10 pages, file size 530481 bytes, and no encryption.

Data-integrity issue:

  • The abstract states that recovery experiments used two concentration levels, 30 and 50 mg kg-1. Rendered Table 2 instead labels the recovery levels as 0.5 ppm Hg and 0.3 ppm Hg, with achieved concentration means 0.471 mg kg-1 and 0.295 mg kg-1.
  • The methods state a total of 375 canned fish samples, with skipjack tuna n=90, bluefin tuna n=95, yellowfin tuna n=90, and albacore tuna n=100. The results narrative later says that, “from the sample of 102 canned fish samples,” no sample exceeded the Turkish Food Codex limit. The source does not reconcile 102 against 375.
  • The abstract states that only 4.8% of investigated samples were detected at a detectable mercury level and that 1.3% exceeded the FAO/WHO 0.5 mg kg-1 recommendation. Table 4 then gives species-level mean/minimum/maximum mercury values for four tuna groups, but it does not state whether these are all samples, detected-only samples, or another subset.
  • The results text says “Table 3 shows results obtained for Hg in 375 samples,” but rendered Table 3 is a certified-reference-material uncertainty table, not a sample-results table. The occurrence table is rendered as Table 4.
  • Table 4 reports, for example, canned skipjack tuna mean 0.381 mg kg-1, standard deviation 0.264 mg kg-1, standard error 0.063 mg kg-1, minimum 0.09 mg kg-1, and maximum 0.66 mg kg-1. The narrative immediately below reports skipjack as 0.381 +/- 0.005 mg kg-1 ww, bluefin as 0.274 +/- 0.005, yellowfin as 0.173 +/- 0.005, and albacore as 0.079 +/- 0.005, which does not match Table 4’s standard deviations or standard errors.
  • The results narrative reports the “smallest Hg” as 0.046 +/- 0.845 mg kg-1 ww and the “highest value of Hg” as 0.382 +/- 0.153 mg kg-1 ww; those summary pairs do not reconcile with Table 4’s minimum/maximum values (<LOQ to 0.66 mg kg-1) or its species means.
  • The discussion states that the highest mercury level determined in the study was 45.33 ug g-1 while also saying detected values were lower than recommended acceptable limits. 45.33 ug g-1 equals 45.33 mg kg-1, which conflicts with Table 4’s maximum 0.66 mg kg-1 and with the claim that levels were below recommended limits.

Disposition:

  • I stopped before writing a wiki/sources/ page for this HJBC candidate, before appending it to data/evidence/audit-queue.csv, before running a source-route audit for it, and before committing any HJBC ingest artifact.
  • Suggested queue disposition: send the PDF to Claude/human audit for whether it can be ingested with exact source-reported values and a prominent contradiction note, whether only method-validation context should be retained, or whether a corrected article/source table exists.