DR02 — Redfish circle-hook/J-hook evidence — v1.0.0 — October 21, 2026
Focused secondary review of three primary studies, not a systematic review, meta-analysis, Louisiana field experiment or exhaustive literature census. Source titles, URLs, table/page locators and PDF hashes are in provenance.json; factual transcription and scope in extraction.json. Study authors collected the fish. The website extracted aggregate counts, computed percentages/intervals and prepared an original comparison plot.

REPRODUCE
Download extraction.json and reproduce.mjs into one folder. With Node 24:
  node reproduce.mjs > reproduced-analysis.json
Output matches analysis.json. No external packages/network required. The module validates counts, calculates n-deep proportions, and keeps mortality unobserved for selectively followed adult arms. Do not replace a blank with zero.
Repository calculation:
  node site/scripts/build-hook-evidence.mjs
Original plot:
  python site/scripts/plot-hook-evidence.py
Plotting runtime: Python, matplotlib 3.11.2. Website builds do not run scientific plotting or download source papers. The chart uses calculated proportions/intervals, not a tracing/adaptation of a source figure.

CALCULATION AND DICTIONARY
n = examined fish in this source analytical arm; not standardized effort or all bites.
deep = fish classified deep-hooked by this source, not observed deaths.
shallow = examined fish classified shallow; n=deep+shallow.
deaths = observed deaths in the stated observation subset, not proof every other capture survived.
mortality_followup_n = number followed for short-term mortality; all subadults, only deep-hooked adults in Charleston. Missing whole-arm mortality is null in JSON and blank CSV.
deep_percent=100*deep/n. mortality_percent=100*deaths/n ONLY when mortality_followup_n=n; otherwise null. Adult whole-arm mortality intentionally not calculated.
Wilson descriptive 95% interval: z=1.959963984540054; p=x/n; denominator=1+z*z/n; center=(p+z*z/(2*n))/denominator; half=z*sqrt(p*(1-p)/n+z*z/(4*n*n))/denominator; bounds=center +/- half. This assumes independent fish; no trip/angler clustering or adjusted effect inference. It does not incorporate longer-term survival, selection, bait/gear or holding effects. No new significance tests or mortality meta-analysis.
Subadult deep-frequency absolute contrast = J percentage minus non-offset circle percentage. Relative reduction =100*(1-(6/58)/(23/57)); not a survival reduction or universally causal geometry effect. Rates rounded only for display.

SOURCE-SPECIFIC BOUNDARIES
Vecchio & Wenner 2007, Table 1 p894: n57 J,58 non-offset,57 offset subadults; deep23,6,11 and deaths4,1,6 respectively. All followed48h. Circle 4/0 versus J2/0; circle offset15 degrees. Natural live mummichogs/cut bait, not a shrimp-cork or lure trial. Methods use L704/L2004ELF; agency outreach lists other model identifiers and broader rates, so we retain the paper-specific factual extraction rather than merge them. Authors' subadult overall-mortality test reports P0.11; failure to detect a difference is not equivalence.
Adults: captured60 J and107 circle; authors excluded2 circle captures because deaths attributed to prior injuries. Audited analytical table: J n60 deep17; circle n105 deep3. Table-based17/60=28.3% differs from abstract's coarse30%; we do not invent18 fish. Only deep-hooked adults held48h. Shallow-death entries N/A; do not infer zero deaths or all-adult survival. Adult plain-shank J7/0 versus circle9/0; bottom longlines, cut mackerel; cannot pool with subadult spinning gear.
Flaherty et al2013: nine live-bait mortality experiments Nov2005–Mar2008;251 held,14 deaths; 1/0 J/non-offset circle. Separate1405 monthly survey and237 survivors in tag analysis. Hook type associated with deep hooking, temperature and position with mortality; authors did not detect direct hook-type mortality association. No hook-specific death rates reconstructed from graphs. Article2013, SEDAR wrapper2016. Standard length differs from total length.
Beckwith & Rand2005:11 Table1 arms1999–2003, deep-hooking only; no mortality follow-up. J comparator is octopus7/0, not Charleston's plain-shank pattern. Short/fixed0.15m/85g; long/sliding0.5m/56.7g. Attachment, mass, leader and year can differ;2003 has no concurrent J control. No universal leader-length, bait or size prescription. Keep source metric values rather than silently correcting approximate/inconsistent inch conversion.

RIGHTS AND REVIEW
Public exports contain attributed factual counts, original arithmetic and original figure, not source PDF text/images or source table layout. Primary PDFs retained only in repository research with source hashes; no blanket redistribution license claimed. Citation links point to original hosts. No partner endorsement or product purchasing recommendation. Independent fisheries/statistical review not performed; single-agent extraction checked visually and computationally, not independently double extracted.
Current handling guidance checked on NOAA Fisheries website October21,2026. Historical papers do not establish current legal rules. A new source/correction requires versioned extraction and preservation of this snapshot. No unmeasured catch, lure, long-term survival or Louisiana local rate inferred.