Claim Ledger — R06: Organic vs. Synthetic/Mineral Fertility in Cannabis Cultivation

Version: v1.0 — signed off by Danny September 7, 2026 (on ChatGPT’s APPROVE WITH MINOR CALIBRATIONS review of v0.2, no new research or correction round; nine wording/boundary edits applied — M4 membrane wording; C024/C025/M6 regime-vs-duration; C013 Mg-only scope; C033 contradiction; C036 archive boundary; C037 confound wording; C038 PGR recollection separated from regulation; C031 ordering qualified to the cited survey; inline same-lab reminder on the Bernstein dose-response cluster) (v0.1 drafted 2026-09-07; v0.2 same day after verify-D on Gemini’s correction round: all 24 replacement DOIs resolve, so the three ⏳ PROVISIONAL identity holds on C024, C025 and C031 are lifted; claim-support on S-209 and S-210 remains abstract-level and their confidence grades are unchanged; Gemini’s corrected export is not used as a source text — see critique §8) · Last updated: September 7, 2026 · Brief: v1.0 FROZEN 2026-09-07 · Protocol: v1.4 Grading: evidence status × cannabis-specificity × confidence per PROTOCOL §3, plus the standing quantitative-claim fields: every load-bearing claim carries an evidence-type tag (measured / hybrid / modeled / observed-administrative / regulatory / absence / methodological) and every quantitative claim carries an inline boundary line — system/medium · functional unit · outcome measure and method · population/genotype — plus R06’s eleventh field: input chemistry + rate + equivalence basis (equal total N / equal EC / equal label rate / equal plant-available N / unmatched), co-inputs, and irrigation-water chemistry (water-undisclosed where the paper does not report it). Three outcomes are never merged: cannabinoid concentration (% or mg/g) · flower yield (g/plant, g/m²) · total cannabinoid yield (mg/plant). Sensory claims name their class: (a) chemistry · (b) trained panel · (c) consumer preference · (d) anecdote. Sources: cited by verified tuple; new registrations S-198–S-218 in SOURCE_LIBRARY (R06 section); 12 further citations reuse existing entries (S-006, S-099–S-104, S-126, S-135, S-161, S-175, S-196). Every citation passed R06’s two-check verification (critique.md, verification/R06-verify-A-gemini.md, R06-verify-B-chatgpt.md). Gemini-only sources (S-209, S-210, S-211, S-212) enter under editor-verified tuples only; verify-D (2026-09-07) confirmed every replacement identifier resolves. Where access is abstract-only (S-209, S-210) the row says so and confidence is capped accordingly. Inheritance: R08-C006 (Massuela), R08-C008 (P ceiling), R08-C022 (assay-dependent compliance), R08-C026 (Rx Green flush), R04-C028 (P leachate), R04-C029 (per-kg reversal), R01-C008/C020 (N asynchrony), R01-C018 (functional-unit reversal), R01-C033 (practices-not-labels), R03-C025 (no safety halo) are inherited and extended, not restated. Rule-12 dependence flags: S-175 and S-104 (Westmoreland/Bugbee lab: two experiments, one lab); S-198/S-199/S-200/S-201/S-203/S-204 (Bernstein lab, Volcani: six experiments, one lab, one or two genotypes); S-102 (also Bernstein lab); S-100/S-101 (Zheng lab, Guelph) with S-103 and S-209 (same lab). “Multiple studies” claims below name the labs.

A. Definitions, taxonomy, and what “organic” legally means

A practitioner-record passage is withheld from the public copy.

R06-C002 — “Organic” and “mineral” programs are chemically overlapping: certified-organic liquid fertilizers deliver soluble ionic fractions and are formulated with chelating agents or added mineral salts, and mineral programs routinely bundle organic chelators and biostimulants. In the one controlled comparison the organic stock was a commercial “Bio NPK 5-2-5” liquid plus a carbon product plus Epsom salt (MgSO₄), and the mineral stock was ammonium sulfate plus commercial soluble fertilizers.

• Evidence status: established · Specificity: direct (one program pair) / general (taxonomy) · Confidence: high · Tag: observed (study inputs) + methodological
• Boundary: input composition as disclosed in S-099 Materials & Methods; no analytical characterization of the organic product’s N speciation reported
• Sources: S-099; ChatGPT A1 taxonomy; Gemini A taxonomy (10 classes); archive 2017 “What Your Plants Want” five-class taxonomy (the author’s archive, practitioner) · Agreement: both
• Notes: Rock phosphate and mined minerals are permitted in organic systems and are mineral materials; “organic ≠ non-mineral.” Where a source does not disclose chemistry, the ledger grades the claim as label-only.

R06-C003 — Four equivalence bases — equal total applied N, equal solution EC, equal label rate, equal plant-available N — are distinct and never interchangeable. The one controlled organic-vs-mineral trial matched treatments on nominal solution concentration (stock 240 N / 96 P / 240 K / 74 Ca / 18.5 Mg mg/L, diluted to 80 / 160 / 240 mg N/L) and still measured different nitrogen uptake efficiency between programs (NUpE 0.55–0.58 organic vs 0.71–0.74 mineral), so nominal equality did not produce equal plant-available N.

• Evidence status: established · Specificity: direct · Confidence: high · Tag: methodological (with measured support)
• Boundary: S-099 Table 6; NUpE per plant over the flowering period; indoor, peat/green-fibre + perlite, one CBD genotype
• Sources: S-099 (full text) · Agreement: both (ChatGPT C2; Gemini “four non-interchangeable equivalence bases”)
• Notes: The brief’s rule, now with its first cannabis datum. Every comparison row below names which basis applied or “unmatched.”

R06-C004 — Cannabis cannot carry the federal USDA “organic” designation because it is ineligible under the National Organic Program; California’s OCal is a state “comparable-to-organic” certification (CDFA for cultivation, CDPH for manufacturing) with its own regulatory meaning, distinct from casual grower use of “organic.”

• Evidence status: established · Specificity: direct · Confidence: high · Tag: regulatory (as of 2026-09-07)
• Sources: S-215 (CDFA OCal page; CDPH OCal page; 3 CCR Div. 8 Ch. 3 OCal Program — chapter verified, section range unconfirmed) · Agreement: both
• Notes: Gemini’s “3 CCR §§10000–10701” range and “OIM forensic isotopic audits” are not carried (unverified/unsupported). S-196 (R04) already registers OCal as a program; S-215 is the primary regulatory citation.

R06-C005 — Cannabis plants and cannabis cultivation are outside the scope of the Canada Organic Regime and cannot bear the Canada Organic Logo under it; any “organic cannabis” claim in Canada rests on a provincial or private certification that must be named.

• Evidence status: established · Specificity: direct · Confidence: high · Tag: regulatory (as of 2026-09-07)
• Sources: S-214 (CFIA organic-products page; wording verified; older CFIA URLs return 410) · Agreement: both (Gemini cited the CGSB standards S-pending; ChatGPT cited CFIA)
• Notes: The CAN/CGSB-32.310/32.311 standards exist (publications.gc.ca); their cannabis-specific application was not checked and is not claimed.

B. Head-to-head evidence and dose-response

Same-lab cluster (rule 12): C009, C010, C012 and C013 (S-198–S-201, S-203, S-204) and the S-102 arm of C011 all come from the Bernstein laboratory (Volcani Center) using one or two genotypes in perlite drip fertigation. They are multiple experiments, not multiple independent research groups. Synthesis must not write “several studies show…” about this cluster without naming the lab.

R06-C006 — Keystone (extends R08-C006). In the only controlled organic-vs-mineral comparison located in drug-type cannabis, mineral fertigation produced 20.9% more inflorescence dry matter than organic fertigation (23.7 vs 19.6 g/plant), organic produced 12% higher CBD concentration (6.5% vs 5.8%), and total CBD per plant remained higher under mineral (1,319 vs 1,197 mg, +10.2%). The three outcomes moved in different directions; the authors attribute the organic program’s lower biomass to lower late-flowering nutrient acquisition.

• Evidence status: probable (single controlled study) · Specificity: direct · Confidence: high · Tag: measured
• Boundary: indoor growth chamber; peat/green-fibre + perlite in containers; one CBD-type genotype; per plant; inflorescence dry matter, CBD % (HPLC), mg CBD/plant; AMF applied to all plants; no leaching; fertilizer twice weekly at nominal 80/160/240 mg N/L. Equivalence basis: matched nominal solution N-P-K-Ca-Mg concentration; unmatched on plant-available N (C003). Co-inputs: AMF common to all; organic program included a carbon product and Epsom salt. Water-undisclosed pending full-text re-read.
• Sources: S-099 (PMC10547009 full text) · Agreement: both; Gemini’s percentages are correct arithmetic on the paper’s values
• Notes: “Organic yields less” as a category law is not what this shows (M3 verdict). Percentages differ from R08’s “≈+21%” and SOURCE_LIBRARY’s caution “17%” only by direction of the ratio (23.7/19.6 = +20.9% mineral over organic; 19.6/23.7 = −17.3% organic under mineral) — both are correct; the ledger uses the mineral-over-organic form.

R06-C007 — In the same trial, the 160 mg N/L rate produced about 95% of the CBD yield of the 240 mg N/L rate while receiving one-third less fertilizer, across both program types.

• Evidence status: probable (single study) · Specificity: direct · Confidence: medium · Tag: measured
• Boundary: as C006; mg CBD/plant; 160 vs 240 mg N/L within program
• Sources: S-099 · Agreement: chatgpt-claim, editor-verified in full text
• Notes: A nutrient-use-efficiency finding, not a footprint finding (no leaching, no LCA).

R06-C008 — Liquid organic fertilizer has a dose-response curve in cannabis: in coir-based substrates, vegetative-stage yield peaked at a rate supplying 389 mg N/L (1.8× the yield at the lowest rate), and flowering-stage optima were substrate-dependent at 212–261 mg N/L in one substrate while the highest rate tested (283 mg N/L) maximized yield in the other; higher rates diluted THC, THCA and CBGA concentrations.

• Evidence status: probable · Specificity: direct · Confidence: medium (HortScience full text paywalled; abstracts verified) · Tag: measured
• Boundary: indoor; two coir-based substrates; single organic liquid program (single-program gradient — unmatched, no mineral arm); per plant; cultivar ‘WP:Med (Wappa)’ in the flowering study; water-undisclosed
• Sources: S-100 (vegetative), S-101 (flowering) · Agreement: both (Gemini’s DOI for S-101 corrected in the correction round; Gemini’s “yield plateau” and ‘OG Kush × Grizzly’ withdrawn)
• Notes: Refutes the categorical form of M5 (“can’t overfeed with organics”) only as far as “organic programs have optima”; no organic-program toxicity trial located (C-M5).

R06-C009 — Mineral nitrogen dose-response in medical cannabis: vegetative development was optimal at 160 mg/L N (deficient at 30, excessive above 160); in flowering, inflorescence yield increased with N up to about 160 mg/L while THCA, CBDA and terpenoid concentrations declined as N supply rose — concentration and productivity diverge.

• Evidence status: probable · Specificity: direct · Confidence: medium (S-199 paywalled; its findings restated in the open-access S-200) · Tag: measured
• Boundary: greenhouse; perlite, drip fertigation; ‘Annapurna’ (THC-type); N supplied as 80% NO₃⁻ / 20% NH₄⁺ (S-198); per plant; single-program mineral gradient (unmatched); water-undisclosed
• Sources: S-198 (vegetative), S-199 (flowering; via S-200 restatement) · Agreement: both (Gemini’s “≥90% NO₃” corrected to 80/20; “linear” decline not adjudicable)
• Notes: Anderson 2021 (C016) found much lower optimal rates in essential-oil hemp — genotype/system dependence, not contradiction.

R06-C010 — Nitrogen form matters independently of total N: at a fixed 200 mg/L N, raising the ammonium fraction to 50% cut inflorescence yield by 46% relative to nitrate-only nutrition, with foliar toxicity symptoms; 100% NH₄⁺ caused plant death; 10–30% NH₄⁺ carried smaller penalties. Cannabinoid and terpenoid concentrations also fell with rising NH₄⁺.

• Evidence status: probable (single controlled study) · Specificity: direct · Confidence: high · Tag: measured
• Boundary: greenhouse; perlite drip; ‘Annapurna’; NH₄⁺ at 0/10/30/50/100% of 200 mg/L N; per plant; equal total N across treatments; roots developed normally at 0–50% NH₄⁺ (necrosis was foliar); water-undisclosed
• Sources: S-200 (PMC9198551 full text) · Agreement: both (Gemini’s “160 mg/L,” “root necrosis” and “pH <4.5” corrected/withdrawn)
• Notes: M4 verdict lives here: NO₃⁻ and NH₄⁺ are different species handled by different transport and assimilation machinery; which species, in what ratio, at what rhizosphere pH, is agronomically decisive.

R06-C011 — Extends R08-C008. Phosphorus above sufficiency does not raise yield or cannabinoid output in cannabis: 30 mg/L P gave ~80% of maximum yield and P above 5 mg/L reduced THCA/CBDA by up to 25% (S-102); raising P from 25 to 75 mg/L gave no yield or CBD gain while leachate P rose 12-fold and P-use efficiency fell from 80% to 31% (S-175); raising P from 15 to 90 mg/L and doubling solution EC from 2 to 4 mS/cm gave no yield or quality gain in closed hydroponics (S-104); a response-surface model placed the flowering optimum at 59 mg/L P and 194 mg/L N (S-103). The 30-vs-59 optimum disagreement is preserved.

• Evidence status: established (direction) / disputed (optimum) · Specificity: direct · Confidence: high · Tag: measured (S-102, S-175, S-104) + modeled (S-103)
• Boundary: S-102 greenhouse perlite, two genotypes; S-175 peat-perlite containers, high-CBD, leachate P per growing area; S-104 closed recirculating hydroponics, medical cannabis; S-103 deep-water culture (not coir/peat — Gemini’s boundary corrected), high-THC cultivar. All single-program mineral gradients (unmatched). Rule 12: S-175 and S-104 are one lab; S-102 is one lab; S-103 is a third lab — three labs, four experiments.
• Sources: S-102, S-175, S-104, S-103 · Agreement: both
• Notes: M9 verdict. No archive P concentration exists to test the archive’s 1-4-5 “bloom” ratio against (archive-audit claim 4).

R06-C012 — Potassium in medical cannabis: 15 mg/L K was deficient, 60–175 mg/L supported high function and yield, 240 mg/L produced oversupply symptoms in one of two cultivars, and most cannabinoid and terpenoid concentrations declined as K rose; separately, K had no effect on inflorescence yield across 60–340 mg/L in a deep-water-culture response-surface study.

• Evidence status: probable · Specificity: direct · Confidence: medium (S-201 abstract-level) · Tag: measured (S-201) + modeled (S-103)
• Boundary: S-201 greenhouse perlite, two cultivars, K 15/60/100/175/240 mg/L; S-103 DWC, 60–340 mg/L; single-program gradients; water-undisclosed
• Sources: S-201, S-103 · Agreement: chatgpt-claim (S-201) + both (S-103)

R06-C013 — Magnesium response is non-monotonic: at the vegetative stage 35–70 mg/L Mg was optimal, 2–20 mg/L impaired development, and 140 mg/L caused only a small physiological decrease with no biomass penalty; at the reproductive stage 35 mg/L was favorable and high Mg reduced reproductive biomass (≈12%). “More magnesium is better” is not supported. (Calcium dose-response was not located; “Cal-Mag” bundles two nutrients and this row speaks only to Mg.)

• Evidence status: probable · Specificity: direct · Confidence: medium · Tag: measured
• Boundary: greenhouse perlite; medical cannabis; Mg 2/20/35/70/140 mg/L; vegetative (S-203) and reproductive (S-204) experiments kept separate; water-undisclosed
• Sources: S-203 (PMC10386616 full text), S-204 · Agreement: chatgpt-claim, editor-verified (ChatGPT’s “impairing at 140” corrected to the vegetative wording)
• Notes: Practitioner precedent: RO water + routine Cal-Mag (archive-audit answer 5) — a water-treatment interaction, not a dose finding.

R06-C014 — Single-element deficiency reduces floral yield sharply for most nutrients: in hydroponic cannabis, floral yield fell 33–72% under every deficiency tested except −Fe and −Mn. Only iron and manganese were tested among micronutrients; the finding does not extend to Zn, B, Cu or Mo.

• Evidence status: probable · Specificity: direct · Confidence: medium · Tag: measured
• Boundary: hydroponic; single-element withholding; per plant floral yield; one cultivar
• Sources: S-208 (Llewellyn, Golem, Jones & Zheng 2023 — the paper ChatGPT attributed to “Cockson et al.”; Cockson 2019 = S-218, vegetative symptomology only) · Agreement: chatgpt-claim, editor re-anchored
• Notes: Deficiency-diagnosis evidence, not a program-superiority test; R11 handoff.

R06-C015 — Augmenting P, K and Fe above a standard regime in medicinal cannabis did not significantly increase final inflorescence mass or cannabinoid concentration; the fertigation architecture (recirculating vs drain-to-waste) did not conclusively change inflorescence mass but did significantly raise total THC yield under recirculation.

• Evidence status: probable (single study) · Specificity: direct · Confidence: medium · Tag: measured
• Boundary: indoor; medicinal cannabis; per plant; total THC yield vs inflorescence mass kept separate; water-undisclosed
• Sources: S-205 (PMC10847352 full text) · Agreement: chatgpt-claim, editor-verified (wording corrected: “affected yield” → “affected total THC yield”)
• Notes: Delivery architecture is an independent variable (R07 handoff).

R06-C016 — In five essential-oil hemp cultivars fertigated with a soluble 20-20-20 fertilizer from 0 to 600 ppm N, growth, biomass and cannabinoid profiles showed strong genotype × rate interactions, and high fertilizer rate/EC reduced performance under those conditions.

• Evidence status: probable · Specificity: adjacent (essential-oil hemp) · Confidence: medium · Tag: measured
• Boundary: greenhouse; five cultivars; single-program mineral gradient; hemp objective differs from drug-type flower
• Sources: S-206 · Agreement: chatgpt-only
• Notes: The genotype × fertility warning’s clearest datum; optimal rates far below C009’s — system dependence.

R06-C017 — In field hemp, some hybrid programs (compost plus half-rate mineral N) matched or exceeded a full mineral NPK control for biomass and phytochemical yield.

• Evidence status: probable · Specificity: adjacent (field hemp) · Confidence: low–medium (abstract-level) · Tag: measured
• Boundary: open field; hemp; per area; equivalence basis: reduced mineral N + compost vs full mineral (unmatched on total N by design)
• Sources: S-207 · Agreement: chatgpt-only
• Notes: Cannot establish indoor drug-type economics or outcomes (C037).

C. Nutrient availability dynamics

R06-C018 — The mechanism proposed for the organic program’s lower late-flower biomass in C006 — mineralization-dependent supply failing to match sink demand when inflorescences become the main sink — is directly discussed and consistent with the measured NUpE gap, but rests on one experiment with one organic product; it is the cannabis instance of the general N-asynchrony mechanism (R01-C008/C020).

• Evidence status: probable (mechanism) · Specificity: direct (one study) / general (mechanism) · Confidence: medium · Tag: measured inputs → inferred mechanism
• Sources: S-099; R01-C008, R01-C020 · Agreement: both
• Notes: Graded separately from the outcome (C006), per the brief. Stress is not assumed as the mechanism for the concentration rise.

R06-C019 — No cannabis-specific mineralization or nutrient-release curves for protein meals, manures, guanos, composts or hydrolysates in containers or beds were located within the documented search (cutoff 2026-09-07); release dynamics in cannabis systems are inferred from general horticulture only.

• Evidence status: unknown · Specificity: direct (absence) · Confidence: high (that none was located) · Tag: absence
• Sources: both search-stopping rationales · Agreement: both
• Notes: The archive’s entire account was “become available when watered” (archive-audit gap). R13–R16 handoff.

R06-C020 — No cannabis isotope-tracer study demonstrating nutritionally meaningful intact amino-acid or peptide uptake from commercial amino-acid fertilizer products was located; plant amino-acid transporters are documented in general plant science, and the quantitative share of N acquired as intact organic N is unquantified in the general literature Gemini cited.

• Evidence status: unknown · Specificity: direct (absence) / general (transporters) · Confidence: high · Tag: absence
• Sources: both searches; S-217 (Näsholm et al. 2009 — states quantification is lacking) · Agreement: both (Gemini’s “<1–5% of N flux” withdrawn as unsourced)
• Notes: M4’s “kernel of truth” and its limit in one line.

R06-C021 — Solution EC is an operational control, not a nutrient-equivalence measure: doubling solution EC from 2 to 4 mS/cm at matched P produced no yield or quality gain (S-104), and high fertilizer rate/EC reduced performance in hemp (S-206), while EC says nothing about N form (C010) or availability schedule (C003).

• Evidence status: established · Specificity: direct · Confidence: high · Tag: methodological (with measured support)
• Sources: S-104, S-206, S-200, S-099 · Agreement: both
• Notes: The archive’s PPM bands (200–500 / 500–1000 / 800–1800 / flush 0–200) are traditional-practice and are not equivalence statements.

D. Quality, sensory, and the flush question

R06-C022 — No peer-reviewed blinded sensory trial (trained panel or consumer) that isolates organic vs mineral fertility while controlling genotype, environment, harvest, drying and cure was located within the documented search. Manufacturer material reporting terpene-profile differences (CANNA/BioCanna) lacks the methods needed for a load-bearing claim.

• Evidence status: unknown · Specificity: direct (absence) · Confidence: high · Tag: absence
• Sources: both searches · Agreement: both
• Notes: M1 verdict. Sensory class (a) chemistry ≠ class (b)/(c) preference (C023).

R06-C023 — Fertility regime measurably changes flower chemistry (sensory class a): terpenoid and cannabinoid concentrations declined as N (S-199/S-200) and K (S-201) rose, and CBD concentration was higher under the organic program in C006 — but no study links those chemistry differences to a trained-panel or consumer sensory outcome.

• Evidence status: probable (chemistry) / unknown (sensory) · Specificity: direct · Confidence: medium · Tag: measured (chemistry) + absence (sensory)
• Sources: S-199, S-200, S-201, S-099 · Agreement: both
• Notes: Gemini’s attribution of “terpene profiles governed by genetics, not fertilizer” to Bernstein 2019 (S-202) was withdrawn — that paper has no organic arm and no terpene data.

R06-C024 — Extends R08-C026. Short pre-harvest water-only flushing (7–14 days) has not been shown to change flower mineral content, yield, potency, terpenes, taste, smoothness or ash color: the one industry trial found no significant differences across 0/7/10/14-day flushes (with flower Fe and Zn rising), a blind expert panel trended toward preferring the unflushed flower (36% rated the 0-day flush “smooth”; not significant), and a Guelph MSc thesis abstract reports flushing “ineffective in removing any significant amount of nutrient from the bud.”

• Evidence status: probable · Specificity: direct · Confidence: medium (industry trial + thesis abstract; no peer-reviewed short-flush trial) · Tag: measured (industry) + measured (thesis, abstract-level)
• Boundary: S-135: indoor, coco coir, ‘Cherry Diesel’, 0/7/10/14-day water-only flush, panel size unstated (Gemini’s “n=36” was a misreading of 36%); S-209: containerized peat/coir, durations and elements not in the abstract; water-undisclosed both. Practice = pre-harvest nutrient withdrawal; outcome = flower tissue minerals and sensory class (c).
• Sources: S-135, S-209 · Agreement: both (S-135); gemini-only (S-209; identity verified, abstract-level — only the abstract’s wording is carried)
• Notes: M6 verdict, short water-only-withdrawal regime. The disagreement with C025 is regime-dependent, not merely duration-dependent: C024 is short water-only nutrient withdrawal; C025 is prolonged leaching to a target leachate EC ≤0.2 mS/cm — duration has not been isolated experimentally. Independence: S-209 is from the Zheng lab (Guelph) and S-135 is a vendor trial — two independent sources, neither peer-reviewed.

R06-C025 — Extended pre-harvest leaching (0–4 weeks, to leachate EC ≤0.2 mS/cm) in a peer-reviewed greenhouse study reduced inflorescence N, P and K concentrations as flush duration increased, reduced inflorescence dry mass, left THC unchanged, and raised CBD concentration in one of two high-CBD cultivars. Evidence therefore differs by flushing/leaching regime: 7–14-day water-only withdrawal has not demonstrated nutrient removal (C024), whereas this separate 2–4-week leaching-to-low-EC study reported declining tissue N/P/K alongside yield loss.

• Evidence status: probable (abstract-level; single study) · Specificity: direct · Confidence: low–medium until full text is read · Tag: measured
• Boundary: greenhouse; peat/coir containers; ‘THM Jack’ and ‘Southern OG’ (high-CBD); 0–3 and 0–4 week flush regimes; per m² and per plant; tap water; Gemini’s “−0.02 kg/m² per week” and “~15 L per 2.5 L pot” are unverified (not in the abstract)
• Sources: S-210 (Alden & Faust 2025, HortScience 60(10):1818–1825, DOI 10.21273/HORTSCI18752-25 — Gemini’s original DOI was fabricated; the paper is real; replacement DOI confirmed in verify-D) · Agreement: gemini-only, editor-verified identity; claim support from abstract (full text paywalled — magnitudes stay out of synthesis until read)
• Notes: M6 verdict, prolonged-leaching regime: disputed by regime, not refuted. Full text required before any magnitude enters synthesis. The most important new finding of the cycle (critique §2b-3).

R06-C026 — No cannabis evidence was located that white ash indicates low fertilizer residue or a “clean flush”; the one cannabis trial measuring ash color found no flush effect (S-135), and the combustion chemistry offered in support (K/Ca catalysis of char burnout, from tobacco and biomass literature) is general, not cannabis-specific, and was not retrievable for verification.

• Evidence status: unknown (unsupported) · Specificity: direct (absence) / general (combustion) · Confidence: high (that no cannabis evidence exists) · Tag: absence
• Sources: S-135; both searches; Gemini’s Garner 1907 / Baker 1981 (S-pending, unverified content) · Agreement: both (Gemini downgraded REFUTED → UNSUPPORTED in the correction round)
• Notes: M2 verdict. “Unsupported” ≠ “ash color can never correlate with anything.”

A practitioner-record passage is withheld from the public copy.

R06-C028 — No study was located linking fertility program to nitrate, nitrite or other combustion-relevant residues in finished flower and onward to smoke chemistry or harshness.

• Evidence status: unknown · Specificity: direct (absence) · Confidence: high · Tag: absence
• Sources: both searches · Agreement: both
• Notes: The 2016 archive claim “you’d be smoking nitrogen… harsh” (archive-audit claim 1) has no evidence base in either direction.

E. Safety, contamination, and compliance

R06-C029 — Canada’s regulator (CFIA) identifies fertilizers and supplements used in cannabis cultivation as potential pathways for heavy metals, pathogens, pesticide residues, dioxins and furans, and notes that cannabis can take up metal contaminants from nutrient solutions or foliar inputs — applying to inputs regardless of organic or conventional status.

• Evidence status: established · Specificity: direct · Confidence: high · Tag: regulatory (as of 2026-09-07)
• Sources: S-213 · Agreement: both

R06-C030 — Extends R03-C025. Contamination patterns cut across market categories: in the Health Canada–led comparison, illegal dried cannabis carried more As, Cd, Pb and Hg, while legal product carried about threefold more chromium; the authors propose fertilizer or micronutrient additives as a possible chromium source but did not test it. No cannabis study traces a known fertilizer metal load through uptake into compliant or failing flower.

• Evidence status: established (pattern) / unknown (fertilizer cause) · Specificity: direct · Confidence: high · Tag: observed (laboratory) + absence (transfer)
• Boundary: 50 legal + 50 illegal Canadian samples; enforcement-seizure selection on the illegal side (R03 caveat)
• Sources: S-161 (full text) · Agreement: both (ChatGPT’s “no fertilizer causation” corrected to “proposed, not tested”)
• Notes: M7 verdict: no safety halo for organic or mineral.

R06-C031 — In one general soil-science survey of fertilizers and amendments, trace-element concentrations generally decreased in the order rock phosphate > sewage sludge > manufactured P fertilizer > organic amendments (the ordering is that survey’s summary across the elements it analyzed and has not been checked element by element; it is not asserted as a universal hierarchy), and commercial-grade P fertilizers raised seed Cd relative to reagent-grade P in flax — but no cannabis measurement of input Cd/As/Pb through substrate to flower exists, so input contamination and flower contamination remain separate claims.

• Evidence status: probable (general ordering) / unknown (cannabis transfer) · Specificity: general · Confidence: medium · Tag: measured (general) + absence (cannabis)
• Sources: S-211 (Raven & Loeppert 1997), S-212 (Jiao, Grant & Bailey 2004 — Gemini’s “2012” tuple was fabricated; both replacement DOIs confirmed in verify-D) · Agreement: gemini-topic, editor re-anchored; Gemini’s “5–150 mg/kg Cd,” “TF 0.3–0.8” and “primary vector of batch failures” withdrawn as unsourced
• Notes: The EDTA/DTPA metal-mobilization claim (Gemini C018) is unresolved (Ahmad 2017 unretrievable) and is not carried.

R06-C032 — Inherits R08-C022. Microbial compliance outcomes depend on assay (culture-based TYMC vs pathogen-specific qPCR); no controlled trial was located showing that an organic fertility program causes a specific pathogen failure rate, and the pathology literature attributes microbial-threshold failures to both conventional and organic methods.

• Evidence status: probable (assay dependence) / unknown (organic-specific rates) · Specificity: direct · Confidence: medium · Tag: measured (S-126) + absence
• Sources: S-126, S-216 (Punja 2021, Pest Manag Sci 77:3857 — Gemini’s venue/DOI corrected) · Agreement: both (Gemini’s organic-specific generalization withdrawn)

F. Environmental fate and footprint of the fertility choice

R06-C033 — Inherits R04-C028. The located measured nutrient-loss evidence in cannabis/hemp is limited to the phosphorus leachate trial (12-fold leachate P for 3× input, no yield gain); no comparable cannabis study of N or K loss under any fertility program, no field or facility P-loss study beyond that chamber trial, and no measured flush-discharge load study was located.

• Evidence status: probable (P) / unknown (N, K, discharge) · Specificity: direct · Confidence: high · Tag: measured + absence
• Sources: S-175; R04-C028; both searches · Agreement: both

R06-C034 — No cannabis LCA matched on functional unit compares organic, mineral and hybrid fertility programs; environmental rankings of fertility programs can reverse between per-area and per-product framings (R01-C018, R04-C029), and a per-kg penalty derived from a yield gap alone is not a footprint (Gemini’s “+26.5% per kg” was retired: co-anchored on a nonexistent paper, arithmetically wrong, and omitting every non-yield term).

• Evidence status: unknown (matched LCA) / established (reversal principle) · Specificity: direct (absence) / general (principle) · Confidence: high · Tag: absence + methodological
• Sources: R01-C018, R04-C029; both searches · Agreement: both
• Notes: M10 verdict stays open, as in R04.

G. Practice, economics, and adoption

R06-C035 — No dataset comparing organic, mineral or hybrid programs on cost per unit output, labor, consistency, storability or failure rate in regulated cannabis was located (searched and empty in both passes; echoes R08-C029).

• Evidence status: unknown · Specificity: direct (absence) · Confidence: high · Tag: absence
• Sources: both search-stopping rationales; R08-C029 · Agreement: both

A practitioner-record passage is withheld from the public copy.

A practitioner-record passage is withheld from the public copy.

Practitioner-record section withheld from the public copy.

Pre-registered public claims — ledger verdicts (M1–M10)

Named data gaps (searched, none located — cutoff 2026-09-07)

1. Replicated multi-genotype organic-vs-mineral trials with analytically characterized inputs and equal plant-available N. 2. Cannabis mineralization curves for organic inputs (C019). 3. Co-input isolation trials (C036). 4. Peer-reviewed blinded sensory testing by fertility (C022). 5. Peer-reviewed short-flush trials with tissue chemistry, combustion products and blinded sensory (C024); full text of S-210. 6. Fertilizer-to-flower metal transfer with known input loads (C031). 7. Amino-acid tracer studies in cannabis (C020). 8. Nitrate/nitrite-to-harshness (C028). 9. Economics on per-kg or per-g-cannabinoid units (C035). 10. Matched LCAs by fertility category (C034). 11. Irrigation-water chemistry reported in fertility trials — none of the located trials was reconstructed for it; all rows above carry water-undisclosed pending full-text re-reads. 12. Copper residue from organic-approved fungicides on flower (brief E13) — not addressed by either pass; open, not “searched and empty.”

Retired / not carried from raw inputs

• Gemini C013 (+26.5% per kg) — retired (nonexistent co-anchor; wrong arithmetic). — Gemini C015 “<1–5% of N flux” — withdrawn. — Gemini C012/C018 numbers — withdrawn. — Gemini C011 as hybrid — reclassified general, folded into C026. — Gemini’s “Bernstein 2019 terpene/organic” attribution — withdrawn. — ChatGPT “Cockson et al.” — re-anchored to S-208.

Downstream handoffs

• R07 (soil vs coco/hydro): C015 (fertigation architecture), C021 (EC), C025 (leaching regime), C032 (assay).
• R11 (nutrition & environmental responses): C009–C014 dose-response cluster; C014 deficiency.
• R12 (cannabinoids & terpenes): C023 (concentration vs yield; terpenoid decline with N/K).
• R13–R16 (soil ecology): C019, C020, C036.
• R17–R20 (inputs & IPM): C002 taxonomy; C029–C031 input safety; copper gap.
• R21 (home grow) / Grow Your 6: C024–C027 (Ch 8 flush reversal — now duration-split), C036–C038 (Ch 1 and Ch 6), C013 (Cal-Mag), C011 (bloom boosters).
• R08 patch candidate (rule 10): R08-C026’s “Stemler” note → “Stemeroff” (S-209) after a second check.