1. Executive Summary

A cannabis flower can look completely clean and still carry material a grower cannot see. That is the central, evidence-supported finding of this paper, and everything else in it builds toward or qualifies that one sentence.

Peer-reviewed research on cannabis flower — concentrated in a small number of research groups (notably Zafar Punja's lab at Simon Fraser University and Kevin McKernan's group at Medicinal Genomics), with a growing body of independent work from Italy, the U.S. Northeast, and the Pacific Northwest — repeatedly documents fungal organisms (Aspergillus, Penicillium, Cladosporium, Fusarium, Botrytis), yeasts, and, with more contested evidence, bacteria present on flower that showed no outward sign of contamination. Some of this material lives inside plant tissue as an endophyte, genuinely undetectable without dissection or lab analysis. Some of it is fine enough — a single fungal spore, a mite fragment, a dust particle — to fall far below the resolution of the human eye. Cannabis regulators use laboratory testing, rather than visual inspection alone, to assess microbial compliance in jurisdictions that regulate microbial contaminants, because visual inspection cannot establish microbiological cleanliness.

At the same time, the evidence does not support treating cannabis as uniformly or dangerously contaminated. Detection is not viability, viability is not toxin production, and toxin production is not demonstrated harm — these are four different claims, and cannabis research repeatedly shows they diverge. The only well-documented case of mass human illness tied to cannabis contamination is a 1981–82 Salmonella outbreak involving grossly elevated bacterial loads traced to a specific manure-contamination event, not routine handling. Human harm from fungal contamination is otherwise limited to case reports of invasive aspergillosis, almost exclusively in immunocompromised patients. No source in this paper supports a claim that ordinary, regulation-compliant cannabis flower poses a demonstrated health risk to a healthy consumer.

This paper finds real, cannabis-specific evidence that harvest and post-harvest handling — not just field or grow-room conditions — measurably raise microbial load; that indoor cultivation does not guarantee sterile flower; that outdoor cultivation carries a different, not simply worse, contamination profile; and that drying speed and final moisture content are the single most consequential levers a grower controls after the plant is cut. Wildfire smoke and ash exposure in West Coast growing regions has increased sharply over the last two decades, and California growers already respond to this by washing particulate off plants in the field — the closest thing this paper has to a documented environmental change with a documented grower response.

A parallel and equally important line of evidence concerns the general human behavior this paper is built to explain: people routinely clean things that are not visibly dirty, because visual inspection was never the actual test for microbiological or surface cleanliness. That principle is well-established in food safety and hygiene science, and produce — particularly leafy greens and herbs, which are structurally closer to cannabis flower than smooth-skinned fruit — is the strongest available agricultural analogy: routinely washed commercially and, per FDA and USDA guidance, still rinsed by consumers regardless of appearance. Cannabis flower's own surface has genuine, peer-reviewed physical characteristics — sticky, resin-fused trichome heads and a structurally complex, multi-layered inflorescence — that make "I don't see anything wrong" a weaker basis for confidence than growers may assume. What the evidence does not support is the claim that cannabis trichomes function as "flypaper" for airborne debris; no study, cannabis-specific or otherwise, has directly shown particulate adhering to a cannabis trichome. That claim is plausible, not proven, and this paper treats it that way throughout.

This paper also finds, and reports honestly, a significant gap: there is no cannabis-specific, peer-reviewed study of what happens when harvested flower is washed with water or a mild aqueous solution. Every rigorous cannabis decontamination study uses irradiation, ozone, or plasma, not washing. The common grower claim that "cannabinoids are hydrophobic, so gentle washing doesn't remove the resin" fails on its own logic: hydrophobicity is well-documented and genuinely rules out meaningful chemical dissolution of THC and CBD in water, but the same peer-reviewed cannabis literature that documents trichome stickiness also documents that trichome heads physically detach under mechanical stress, and an entire extraction industry (ice-water hash) exists because cold water plus agitation is a proven method for knocking trichome heads loose. Dissolution and mechanical loss are different mechanisms, and no cannabis-specific study has measured either one for washing specifically. Anything said about washing cannabis in this paper has to be built from produce-washing science and cannabis chemistry, clearly labeled as inference, not from a cannabis-specific washing trial that does not yet exist.

Finally, this paper distinguishes prevention from remediation, and finds real evidence for treating them differently. Remediation technologies — chiefly gamma, electron-beam, and X-ray irradiation — are wellestablished, mechanistically sound, and demonstrably effective at reducing microbial load without measurably affecting THC or CBD content. But a 2025 peer-reviewed study found that gamma-irradiated cannabis, including commercial product that passed standard compliance testing, can still carry detectable mycotoxins and residual fungal genetic material — irradiation reliably kills organisms, but it does not reliably undo the toxins those organisms produced before treatment. That is the evidence-backed basis for treating prevention as the stronger long-term position: not because remediation is ineffective or unsafe, but because it is applied after contamination has already occurred, and some of what happens before treatment cannot always be fully undone by it. An ounce of prevention beats a pound of radiation — not because the pound of radiation doesn't work, but because prevention avoids a problem that radiation cannot always fully undo, even when it works exactly as designed.

2. Research Question and Scope

Central question: A cannabis flower can look clean. What does the evidence show may actually be present on it, and what does that mean for post-harvest cleanliness?

This paper establishes, using the strongest available evidence, what visible and non-visible foreign material, fungal organisms, microbial contamination, environmental material, insect-associated material, and handling-related contamination can be present on cannabis flower at harvest and during early post-harvest handling — with particular attention to the gap between what a grower can see and what may be present without being obvious to the naked eye. It also establishes the behavioral and agricultural logic for treating that gap as a reason to clean, examines what is and is not known about washing as an intervention, and distinguishes prevention-stage strategy from remediation-stage technology.

This paper is an evidence review, not a product-efficacy study. It does not evaluate the efficacy of Terps USA Bud Wash or any other specific washing product. Where commercial technologies are discussed, they are evaluated as technology categories using published evidence.

Four categories of evidence are used throughout, and each claim in this paper is explicitly typed as one of them:

• Cannabis-specific evidence — findings from studies conducted directly on cannabis or hemp flower, tissue, or trichomes. This is the strongest evidence category and is prioritized wherever it exists.
• Cross-crop inference — findings from other plants or agricultural commodities (produce, hops, tobacco, grain) applied to cannabis by structural, chemical, or mechanistic analogy. Labeled explicitly wherever used, with the specific points where the analogy breaks down stated plainly.

• General hygiene/agricultural analogy — findings from food safety, public health, or consumerbehavior research that establish a general principle (e.g., visual cleanliness is not microbiological cleanliness) rather than a cannabis-specific fact.
• Product-efficacy evidence — evidence about what a specific commercial product or technology does. This paper does not evaluate Bud Wash under this category; it does evaluate published research on irradiation, ozone, and controlled-environment drying as technology categories, not specific commercial products, per Sections 14–15.

Genuine gaps are stated as gaps, not inferred around. The largest of these — the complete absence of a cannabis-specific washing study — is treated as the paper's most important open question, not minimized.

3. What Can Be Present on Cannabis Flower at Harvest?

Drawing on Sections 5–7, the credible evidence base supports that cannabis flower at harvest and during early post-harvest handling can carry, in varying and study-dependent proportions:

• Fungal organisms and spores, both epiphytic (surface) and endophytic (internal, symptomless) — genera repeatedly and independently documented include Aspergillus, Penicillium, Cladosporium, Fusarium, Botrytis, and powdery-mildew fungi (Golovinomyces spp.) (Gwinn et al. 2023; McKernan et al. 2016; Punja et al. 2019, 2023).
• Yeasts and, with more contested evidence, bacteria, at levels that vary enormously by genotype, environment, and handling practice (Punja et al. 2023; Comeau et al. 2020).
• Environmental particulate — pollen (well-documented), wildfire ash and smoke residue (welldocumented in outdoor cultivation and now further supported by the environmental-trend evidence in Section 12), and, by inference from other crops, dust, airborne particulate, and soil/mineral matter (Olejar & Park 2022; Feder et al. 2021; Martin et al. 2025).
• Insects and insect-associated material, most directly documented for corn earworm larvae and frass, hemp russet mites, and cannabis aphids; less documented, but plausible, for other common cannabis pests (Britt et al. 2021; Pulkoski & Burrack 2020).
• Handling-related microbial exposure and load during harvest, trimming, and processing; cannabisspecific studies document an association with worker activity, while individual contact points such as hands, gloves, tools, bins, and surfaces are supported mainly by cross-crop food-safety evidence rather than cannabis-specific isolation studies (Green et al. 2018; Punja et al. 2023).

None of this material is unique to cannabis in kind. What is cannabis-specific is the plant's sticky, trichomecovered, structurally complex flower surface (Section 10), its typically manual harvest and trim process, and its post-harvest drying/curing pathway (Section 6) — all factors this paper treats in detail below.

This section does not, and should not be read to, establish a prevalence figure. No source located anywhere in this paper's evidence base supports a generalizable "% of commercial cannabis carries nonvisible contamination" statistic at national or even regional scope. Every percentage cited later in this paper is facility-, study-, or region-specific and should not be extrapolated industry-wide.

4. What You Can See vs. What You May Not See

The single most important organizing distinction in this paper is between material a grower can identify by unaided visual inspection and material that can be present without producing any visible sign.

This table supports, without overstating, the paper's central claim: looking clean does not necessarily establish that flower is microbiologically clean. It is not evidence that any specific batch of flower is contaminated.

5. Fungi, Spores, and Yeasts

Fungal contamination is the best-documented category in this paper, spanning both culture-based (viableorganism) and DNA/PCR-based (detection-only) methods — a distinction maintained strictly throughout, because cannabis research shows the two methods frequently disagree.

Aspergillus. Culture-based recovery of A. ochraceus, A. niger, and A. flavus from dried inflorescences has been reported at licensed cultivation facilities (Gwinn et al. 2023). Metagenomic sequencing has found Aspergillus DNA comprising 18–58% of classified fungal reads in some dispensary samples, yet the same organism "grew poorly in culture," dropping to roughly 1% of recovered colonies — direct evidence that DNA detection and viable/culturable detection diverge substantially for this genus (McKernan et al. 2016). A follow-up molecular-methods paper found that some qPCR assays used for regulatory testing can crossreact with non-regulated Aspergillus species, producing false positives for the four species most jurisdictions specifically test for (McKernan et al. 2022). Separately, viable Aspergillus has been cultured directly from cannabis smoked by immunocompromised patients who subsequently developed invasive aspergillosis, in case reports spanning five decades; a CDC claims-database study found cannabis users roughly 3.5 times more likely to carry a fungal-infection diagnosis code than non-users, while explicitly cautioning that administrative claims data cannot establish causation (Benedict, Thompson & Jackson 2020). Aspergillus contamination is not exclusively a post-harvest phenomenon: Oregon State University Extension documents that Aspergillus can be seed- or clone-borne and introduced via grow-room HVAC or dehumidification failures pre-harvest, even as the same source confirms that the largest measurable increases in Aspergillus prevalence typically occur post-harvest, during drying and curing. Both stages matter; post-harvest moisture control is the more consequential and more controllable lever (Section 6), but grow-room causation should never be categorically excluded.

Penicillium is the most consistently recovered toxigenic genus across independent cannabis studies, by both culture and sequencing (Gwinn et al. 2023; McKernan et al. 2016; Punja et al. 2019). Mechanical trimming has been shown to increase Penicillium recovery from harvested buds compared to pre-harvest levels, attributed to fungal entry through trim-wound damage (Punja et al. 2019) — direct evidence that a specific, common post-harvest handling step changes fungal load.

Cladosporium is repeatedly documented on cannabis inflorescences by culture, though exact prevalence figures found in secondary sources should be independently re-verified against primary tables before public quotation.

Fusarium species have been culture-recovered from cannabis buds, and a 2025 chemical-analysis study of 118 seized cannabis samples found Fusarium-associated mycotoxins (fusarenon-X, diacetoxyscirpenol, and in one sample ochratoxin A) in 16% of samples via mass spectrometry — an indirect but generally stronger signal of fungal activity than DNA detection alone (Leung et al. 2025).

Botrytis cinerea ("bud rot") is the best-documented flower-specific fungal pathogen. A British Columbia government fact sheet states that Botrytis infections "typically begin internally where humidity is highest, progressing outward" before visible grey sporulation appears — direct support for the "hidden until it isn't" pattern this paper is centrally concerned with. A 2025 peer-reviewed review independently confirms fungi, including Botrytis, can reside within plant tissue "without causing discernable symptoms" (Buirs & Punja 2025).

Powdery mildew (Golovinomyces spp.) is confirmed on cannabis by PCR sequencing and pathogenicity testing (Pépin, Punja & Joly 2018). Unlike Botrytis, the same British Columbia fact sheet states that "systemic infections and internal spread of the pathogen have not been demonstrated" for powdery mildew — it is a surface disease, visually obvious once established, and should not be cited as an example of hidden contamination.

A 2021 study cultured 45 viable fungal colonies directly from cured commercial cannabis flower and confirmed their identity by whole-genome sequencing, spanning Aspergillus, Cladosporium, Botrytis, Candida, Penicillium, Rhizopus, and Fusarium — direct, strong evidence that viable fungal organisms are present on finished commercial flower, not just in laboratory models (McKernan et al. 2021).

Yeasts, bacteria, and total microbial load. A large, cannabis-specific, three-year survey (>2,000 samples, six genotypes) found total yeast and mold counts strongly influenced by genotype, environment, and pre- and post-harvest handling practice: harvesting activity, leaf litter, temperature/humidity, and season all raised counts, while fan-driven air circulation, hang-drying whole stems, and reaching a safe final moisture target lowered them (Punja et al. 2023). A parallel hemp study found hot-air drying achieved up to a 2-log reduction in total yeast/mold counts compared to slow ambient-air drying (Baek, Grab & Chen 2025).

Bacterial evidence is the most methodologically contested category in this paper. One metagenomic sequencing study found sequence reads matching several bacterial genera including Acinetobacter, E. coli, Pseudomonas, and Salmonella enterica (McKernan et al. 2016) — DNA-read detection, not confirmed viable organisms. An independent group, using a different sample set and 16S sequencing, could not reliably amplify bacterial DNA from cannabis flower/aerial tissue at all (Comeau et al. 2020). This paper does not resolve that conflict in favor of the more alarming finding; both are reported, and any claim built on bacterial presence should be qualified as contested.

The one unambiguous, well-documented case of demonstrated human harm from bacterial contamination is the 1981–82 multistate Salmonella muenchen outbreak: 85 confirmed illness cases, marijuana samples from patients' homes contaminated at up to 10 million CFU/g, and molecular fingerprinting linking patient and product isolates (Taylor et al. 1982; CDC MMWR 1981). This involved grossly elevated contamination from an unusual source (manure) and is not evidence of routine risk — but it is the clearest proof that cannabis-borne bacterial contamination can cause real, documented illness at sufficient magnitude. A large commercial testing-lab dataset (nearly 59,000 California samples, 2012–2019, trade rather than peer-reviewed) found roughly 14–16% of dried flower samples failed standard bacterial or fungal action levels, attributing most flower bacterial contamination to post-harvest handling rather than field-stage exposure (Martin et al. 2021).

A 2024 recall of Arizona cannabis products for confirmed Aspergillus and Salmonella contamination is a useful, more typical modern contrast: state officials explicitly reported no associated illnesses (CIDRAP 2025).

What the evidence does not support: a specific, generalizable population-level contamination percentage for commercial cannabis nationally; that detected fungal or bacterial DNA represents viable, infectious organisms; or that any documented mycotoxin/pathogen level in this evidence base has been linked to illness in an ordinary, healthy consumer under routine, regulation-compliant conditions.

6. Harvest and Post-Harvest Handling

This is the section with the strongest, most convergent cannabis-specific evidence in this paper, and it anchors the strongest defensible claims Terps USA can make.

Two independent, peer-reviewed, cannabis-specific studies support that harvest and processing activity — separate from cultivation-stage conditions — measurably raises microbial exposure and load. A NIOSH occupational study at an outdoor cannabis farm found that personal air samples collected on workers during harvesting, bud-stripping, and hand-trimming showed higher bacterial and fungal diversity than ambient area air, with Botrytis cinerea the single most common fungal organism detected in worker breathing-zone air, about 59% of fungal reads (Green et al. 2018). The large multi-year total-yeast-and-mold survey (Section 5) independently found that "harvesting activity by workers" was statistically associated with increased microbial counts, alongside leaf litter on the processing floor (Punja et al. 2023).

Neither study decomposes the effect into specific contact points — hands, gloves, scissors, bins, tables, drying racks. No cannabis-specific study isolates those variables individually. This paper draws, explicitly labeled as a cross-crop parallel rather than cannabis evidence, on food-safety research demonstrating that hand and glove contact is a quantifiable, sometimes highly efficient vector for microbial transfer during manual harvest: one controlled study found a single contaminated glove transferred bacteria to up to 98% of sequentially harvested lettuce leaves in laboratory conditions and 74% of harvested heads in field conditions, persisting through at least 20 sequential harvests from one contamination point (Zhao et al. 2021). An earlier study found roughly a 10% hand-to-food bacterial transfer rate with bare hands, dropping to roughly 0.01% with gloved hands (Montville, Chen & Schaffner 2001). Separately, the general microbiological fact that fungal spores resist plain water rinsing — Aspergillus-type conidia carry a hydrophobic rodlet (hydrophobin) surface layer that confers water-repellency — is independently well documented and supports the plausibility, though not a cannabis-specific demonstration, of reused equipment (drying racks, trays) acting as a crosscontamination vector if not chemically decontaminated between batches.

Verdict on the specific claim under test — "a grower can do everything right during cultivation and still introduce or encounter unwanted material during harvest and post-harvest handling" — is rated B (Good), not A. The first half of the claim (handling activity is an independent, documented contributor to microbial load) is directly supported by two cannabis-specific peer-reviewed studies. The second half (this happens regardless of cultivation quality) is a reasonable logical inference — handling-introduced contamination is mechanistically a separate pathway from cultivation-stage contamination — but no study directly controlled for cultivation quality as a variable to test that inference. This claim should not imply a guarantee or a specific magnitude, and should not claim clean cultivation is irrelevant — only that handling is a real, additional, separately-documented pathway.

Drying and moisture control is the single biggest lever a grower controls after harvest. Freshly harvested cannabis is roughly 75–81% water by weight (Baek, Grab & Chen 2025); mold proliferation in the fungal genera most relevant to cannabis (Aspergillus, Penicillium, Fusarium) becomes a live risk in roughly the 13.5–19% moisture range, per foundational FAO grain-storage guidance — a general post-harvest plant-science parallel, not a cannabis-specific colonization study, and any specific single-number moisture threshold cited publicly should reference this range rather than an unqualified figure. Cannabis-specific review literature states that keeping water activity below roughly 0.7 substantially restricts growth of the most common contaminant genera and that improperly dried, humid-stored material favors mycotoxin accumulation (Gwinn et al. 2023; Al Ubeed, Wills & Chandrapala 2022). Hot-air drying (75°C) achieved up to a 2-log CFU/g reduction in total yeast/mold counts compared to slow ambient-air or freeze drying, from drying method alone, with no decontamination step involved — one important precision note: in the same study, the subsequent curing stage did not independently and significantly reduce microbial levels; the effect is attributable specifically to fast/hot-air drying (Baek, Grab & Chen 2025). Mechanical trimming has been shown to increase Penicillium recovery compared to pre-harvest levels, and a large multi-year survey found trimming method, leaf litter, air circulation, and final moisture target were all statistically significant drivers of total yeast and mold counts (Punja et al. 2019, 2023). A three-year Italian field survey found longer ambient-temperature drying correlated with the most heavily contaminated batches (Spampinato et al. 2024).

No cannabis-specific controlled study was found isolating airflow/ventilation as its own variable, separate from temperature, humidity, and duration. Airflow's role is asserted qualitatively in reviews and stated as good handling practice, but has not been experimentally isolated in a cannabis study — this should be labeled as a well-supported mechanistic inference, not a directly tested cannabis finding, when used publicly.

7. Indoor vs. Outdoor Cultivation

The evidence directly contradicts a simple "outdoor = dirty, indoor = clean" assumption, and this paper does not adopt it.

Indoor cultivation does not guarantee sterile flower. Multiple independent cannabis-specific studies document Aspergillus, Penicillium, Fusarium, and in some studies bacterial signal, in indoor and greenhousegrown cannabis, including a case where standard culture-based facility quality-control testing failed to detect toxigenic fungi that molecular testing found present (Gwinn et al. 2023; McKernan et al. 2016). A large indoorfacility survey found that the flower's own internal microclimate humidity (roughly 75%) substantially exceeded the ambient greenhouse humidity the facility actually controlled (roughly 55%) — direct evidence that facility-level environmental control does not eliminate bud-level microclimate risk (Punja et al. 2023). A fatal 2022 occupational asthma case at a Massachusetts indoor cannabis facility, investigated by the CDC, documented measurable endotoxin and particulate exposure during flower grinding even at levels below regulatory occupational limits (CDC MMWR 2023) — an occupational-safety finding, not a direct consumerproduct-safety finding, but a clear demonstration that "indoor" and "contamination-free" are not the same thing.

Outdoor cultivation carries additional, and in some respects different, contamination exposure. The NIOSH outdoor-farm study found Botrytis cinerea as the dominant fungal species in worker breathing-zone air, rising sharply during harvest tasks (Green et al. 2018). A three-year Italian field survey of two outdoor sites found the wetter, more humid site had significantly higher bacterial and yeast/mold counts than the drier site, with E. coli in roughly half of relevant samples and Salmonella in 15%, concentrated in the wettest year (Spampinato et al. 2024). A Florida arthropod survey found outdoor cultivation faces a broader taxonomic range of insect and mollusk pests than indoor cultivation, which faces a narrower but still real set — aphids, mites, thrips, whiteflies (Ahmed, McKenzie & Osborne 2024).

The best available side-by-side dataset found outdoor-grown buds carried a different dominant fungal profile than indoor or greenhouse buds, not simply a higher load (Punja et al. 2019). The evidence best supports "outdoor exposure is real, weather-driven, and compositionally distinct from indoor," not a simple

"outdoor is dirtier" magnitude claim. No large-scale, statistically powered study directly comparing the same genotype grown simultaneously indoor and outdoor was located — this is a genuine gap.

Wildfire ash and smoke are the strongest, most quantified outdoor-specific exposure in this paper's evidence base, and the strongest example of a genuinely changed environmental condition. A peerreviewed, quantified California grower survey found 57% of respondents reported washing particulate off growing plants as a wildfire-mitigation practice, with growers describing a "wet ash" crust capable of causing full crop loss and 45% reporting increased pest/fungal pressure following smoke exposure (Martin et al. 2025). This is corroborated by independent environmental data: California's population-weighted wildfireattributable PM2.5 rose roughly 8-to-9-fold between 2006 and 2018, a genuine, measured physical increase attributed to warming, drought, and fuel accumulation — not improved detection finding old smoke (see Section 12 for the full environmental-trend analysis). This is the paper's cleanest bridge between "the environment has changed" and "growers already respond by washing," and it is treated in full in Section 12.

8. Why We Clean Things That Look Clean

The question this paper is ultimately answering is narrower than "is cannabis dirty": it is why do people routinely clean things they cannot see are dirty, and does that logic transfer to post-harvest cannabis flower?

The underlying hygiene principle is real and well-anchored. USDA's Food Safety and Inspection Service states plainly that "you can't see, smell, or taste harmful bacteria that may cause illness" — a direct, on-point government statement of the exact proposition this section is built around, in a food-safety context. The physical basis is equally solid: common foodborne bacteria (roughly 1–100 microns) and viruses (under 0.5 microns) sit one to three orders of magnitude below the human eye's approximate 0.1 mm resolution limit — the same physics this paper already applies to cannabis-relevant fungal spores in Section 5. A search for an equivalent explicit CDC handwashing statement ("wash your hands even if they look clean") did not turn up that specific phrasing on CDC's current handwashing pages; this is flagged as a gap rather than assumed, and the USDA statement should be used as the anchor citation instead.

Cornell's Produce Safety Alliance draws a sharp, technical distinction relevant here: "you cannot sanitize a dirty surface... not all surfaces can be sanitized, but all surfaces can be cleaned" — visible cleanliness is a precondition for effective sanitation, not proof that sanitation already happened. A concrete illustration of the gap between perceived and actual cleanliness: medical students judged their own lab coats "clean" while the coats carried substantial Staphylococcus aureus loads on culture — a direct demonstration, in a noncannabis but rigorously documented setting, that self-assessed visual cleanliness and actual microbial load can diverge sharply.

This principle is strongly established in handwashing and food-contact-surface hygiene, where the causal chain (invisible pathogen → transfer → illness) is well studied. It gets weaker the further it is stretched: it does not, on its own, establish that any specific unwashed surface is dangerous, nor does it establish a magnitude of risk. It is a reason to investigate a category, not proof that any given flower lot is contaminated, and this paper does not use it that way anywhere else.

This point is worth stating plainly, without dressing it up: the principle is not a novel insight, it is a description of ordinary behavior most people already practice without thinking about it — which is precisely why it is useful. The purpose of this section is behavioral normalization, not a rhetorical trick.

9. Produce as the Primary Agricultural Analogy

Produce is the strongest available agricultural comparison for cannabis flower — considerably stronger than other inhaled agricultural commodities. (Tobacco was examined in detail during this paper's research process as a potential secondary comparison and was found to cut the other way: tobacco is never washed or submerged in water at any standard commercial post-harvest stage, and its one water-adjacent processing step, steam conditioning, exists to restore leaf pliability for mechanical handling, not to sanitize it. Tobacco is a legitimate example of an inhaled crop undergoing serious post-harvest processing — mechanical threshing, air classification, dust extraction, dedicated foreign-material detection — but it should never be used to argue that washing inhaled crops is standard practice, because for tobacco specifically, it is not. This paper therefore does not use tobacco as a primary analogy and mentions it only for this limiting purpose.)

Commercial washing vs. consumer rinsing vs. pathogen-reduction claims are kept strictly separate, per FDA/USDA framing. Commercial washing happens at the packinghouse, typically via dump tanks or flumes, governed under FDA's Produce Safety Rule, which sets a water-quality standard requiring no detectable generic E. coli per 100 mL for water that directly contacts produce. This is a supply-chain-stage process the consumer never sees.

Consumer rinsing is separately and explicitly recommended, with one clear exception. FDA and USDA guidance instructs consumers to rinse fresh, unsealed produce under running water just before eating, cutting, or cooking, regardless of whether the peel is eaten, and is explicit that rinsing "will not eliminate" bacteria "but can reduce" them. This guidance does not condition the recommendation on the produce looking dirty. The one clear exception: pre-washed, sealed "triple-washed" salad greens are the case where university extension sources (Virginia Tech, University of Minnesota) advise consumers not to rewash, because reopening a sealed, already-sanitized product to home tap water can introduce more risk than it removes. This exception is itself useful evidence, because it shows the "wash it again" instinct is not applied blindly — it is applied where the marginal benefit outweighs the marginal risk, which is exactly the calibrated claim this paper needs, not a blanket "always wash" claim.

Plain-water rinsing is documented to reduce surface microbial load by roughly 1.4–2.9 log CFU/g in a classic controlled produce study, and a 2026 scoping review of 47 produce-washing studies found median pesticideresidue reductions of about 30% for plain water up to about 54% for a dilute vinegar soak (Sapers et al. 2006; de Montagnac et al. 2026). Commercial produce washes and acid rinses generally show no significant advantage over plain water for microbial reduction. None of this eliminates internalized contamination or guarantees pathogen-free produce — FDA and USDA guidance is careful to say washing "reduces," not "eliminates." Shared wash water is itself a documented contamination vector if not actively managed: organic-matter buildup progressively consumes disinfectant capacity across successive batches, meaning unmanaged wash water can spread contamination from one batch to the next rather than removing it (Gil et al. 2015).

Leafy greens and herbs are the closer structural analog to cannabis flower, and this is the single most useful sub-finding for bridging produce science to cannabis. Multiple university extension and peerreviewed sources document that leafy greens' folds, crevices, and stomata reduce the effectiveness of sanitizing rinses compared to smooth-skinned produce like apples — meaning the produce category structurally closest to cannabis flower (complex surface, high surface area, layered structure) is also the category where washing is documented as harder, not unnecessary. A 2018 peer-reviewed study on fresh coriander/cilantro found a three-step plain-water rinse was outperformed by dilute potassium permanganate at reducing bacterial load on this structurally complex herb, reinforcing that plain rinsing has real but limited effect on crevice-bearing produce (Subramanya et al. 2018).

Where this analogy breaks down — stated plainly, not hidden:

• Cannabis is inhaled, not ingested — the produce-washing evidence base is built around oral/dietary exposure and pathogen risk; no direct transfer of that risk model to inhalation exposure should be implied.
• Cannabis is dried and cured, not eaten fresh — final moisture content is the dominant driver of postharvest fungal risk (Section 6); any wet process introduces a re-wetting/regrowth risk with no produce equivalent, because produce is not subsequently dried the way cannabis would need to be.
• No wash method — plain water, dilute solution, or otherwise — has ever been tested on cannabis flower specifically. The produce literature establishes what washing can plausibly do to a similarly structured surface; it does not establish what washing does do to cannabis. This is restated with emphasis because it is the largest single evidence gap in this paper (Section 11).
• There is no cannabis-equivalent regulatory framework for wash-water quality or cross-contamination control analogous to FDA's Produce Safety Rule.

10. Cannabis Flower Structure and Surface Characteristics

The question: does cannabis flower have physical or chemical characteristics that plausibly retain surface material, and is "sticky plant" / "flypaper" a defensible communication concept?

Trichome structure and density are directly and repeatedly documented in peer-reviewed cannabis botany, not just cultivation blogs. Cannabis inflorescences carry three glandular trichome types — bulbous, sessilecapitate, and stalked-capitate — with the stalked-capitate type (the principal cannabinoid/terpene producer) described as "particularly abundant on the female reproductive organs" and increasing substantially through flower maturation (Punja, Sutton & Kim 2023; Adal et al. 2025). Density figures run roughly 18–30 stalkedcapitate trichomes per mm² depending on cultivar and sampling stage.

Physical stickiness is directly observed, not inferred, in peer-reviewed microscopy — this is the strongest single finding in this section. Punja, Sutton & Kim (2023) document by microscopy across a developmental time course that "resin secretion causes trichome heads to stick to one another," sometimes producing aggregates of five to six fused glandular heads as resin accumulates. The same study found machine-trimmed samples showed glandular heads "detached or collapsed" while hand-trimmed samples retained largely intact trichomes — direct, cannabis-specific evidence that the resinous layer is both physically tacky and mechanically fragile under handling (this same finding is the anchor for Section 11's mechanical-loss discussion). Cannabinoids and terpenes accumulate in a subcuticular cavity between a hydrophobic waxy cuticle and the underlying hydrophilic cell-wall matrix — real, structural, cell-biology-level chemistry, separate from the tackiness/viscosity question, which is a physical rather than chemical property.

Inflorescence structure is genuinely complex, per peer-reviewed floral-architecture research: the female cannabis inflorescence is built from "condensed higher-order branchlets" with up to seven visible orders of branching, with each female flower enclosed by a bract that "embraced the carpel(s) and the female flower" (Spitzer-Rimon et al. 2019). This directly corroborates, with real developmental-biology citations, what growers describe informally as densely overlapping bracts, calyxes, and pistils. No source directly states that this structure traps particulate — that connection is a reasonable structural inference from documented architecture, not a tested finding.

Direct evidence that particulate actually adheres to cannabis trichomes or resin does not exist. No peerreviewed study — cannabis-specific or otherwise — was located showing, by microscopy, SEM, or any other method, dust, pollen, fungal spores, or insect material physically adhered to a cannabis trichome or resin droplet. Fungal-detection studies (Section 5) document organisms recovered from flower by culture and sequencing but do not image or describe an adhesion mechanism; detection is not the same claim as documented adhesion. Hops (Humulus lupulus), the closest botanical relative in the same family, carries an analogous lupulin gland structure and is separately documented to carry mycotoxin contamination in a large majority of tested samples — but no study ties that contamination specifically to adhesion at the lupulin gland, so this is a whole-plant contamination parallel, not an adhesion-mechanism parallel. Carnivorous/protocarnivorous plant literature (sundew, Roridula) offers real, rigorously measured physics of viscoelastic plant-resin adhesion capturing insects — genuinely useful as proof that plant resin can mechanically trap material — but these are coevolved, specialized trapping systems entirely unrelated to Cannabaceae and are cited, if at all, only as mechanistic reference, never as agricultural evidence.

Rating: C — Plausible. The mechanistic foundation is unusually strong for a "plausible" rating — real, peerreviewed, cannabis-specific documentation of trichome self-adhesion, hydrophobic resin chemistry, and complex floral architecture all exist. But the central testable claim — that this documented stickiness and structure causes external particulate to adhere to and be retained on the flower — has not been directly observed or measured in cannabis or in the closest available analog.

"Flypaper" is not an established scientific description and should not be used as one. It implies a demonstrated, efficient capture mechanism, which is documented for coevolved carnivorous plants but not for cannabis; cannabis resin's documented biological role is metabolite storage, not particulate capture. Cannabis flower may be accurately and defensibly described as sticky, resinous, and structurally complex — with the qualification that retention of foreign material by that structure is inferred, not directly measured.

11. What Washing Research Does and Does Not Establish

This section is explicitly bounded: it is a review of washing research in general and of cannabis chemistry relevant to washing, not an evaluation of Terps USA Bud Wash, and it does not test, mention, or draw conclusions about that specific product.

The central finding of this section is an evidence gap, and it is reported as such rather than minimized. No peer-reviewed, controlled study was located that washed harvested cannabis flower in water or any aqueous solution and measured the resulting effect on microbial load, cannabinoid content, terpene content, or trichome integrity. Every rigorous cannabis-specific decontamination study instead tests non-aqueous methods: gamma or e-beam irradiation, ozone gas, or non-thermal plasma (Schnabel et al. 2026). No cannabis washing study of any kind — cannabis-specific, controlled, peer-reviewed — currently exists.

What produce-washing science supports, by cross-crop inference (Section 9): plain-water rinsing measurably reduces surface microbial load and some pesticide residue on structurally comparable produce (leafy greens, herbs). This is a reasonable basis for treating "washing can plausibly remove physical debris" as a defensible inference — not a demonstrated cannabis finding.

Chemical dissolution vs. mechanical loss — the distinction the common grower claim collapses. THC's water solubility is documented at 2.8 mg/L at 23°C, per WHO Expert Committee on Drug Dependence critical review data; THC's log P (octanol-water partition coefficient) runs approximately 6.9–7.2. CBD is reported qualitatively as "practically insoluble" in most primary sources, with a similarly high log P. Terpenes are also poorly water-soluble but separately volatile, meaning they can be lost to evaporation — a third loss pathway distinct from both dissolution and mechanical detachment, one that dissolution-only arguments ignore entirely.

This chemistry is genuinely strong, and it genuinely does not answer the question growers use it to answer. The same peer-reviewed microscopy already cited in Section 10 (Punja, Sutton & Kim 2023) documents a structural weak point at the stalk-to-head junction of the trichome and shows machine-trimmed flower with visibly more detached or collapsed glandular heads than hand-trimmed flower — real, cannabis-specific evidence that trichome heads separate from the plant under mechanical stress, independent of any chemistry. More directly relevant, peer-reviewed extraction literature describes conventional ice-water hash production as immersing cannabis in ice water and using mechanical stirring/agitation to separate trichomes from plant material (MacGowan & Martynenko 2025; see also broader extraction reviews). This does not reproduce a gentle post-harvest wash protocol, and it should not be treated as one. It does, however, demonstrate the narrower point that water exposure combined with sufficient mechanical agitation can physically separate trichomes — a mechanism distinct from chemical dissolution and therefore relevant to evaluating any blanket claim that hydrophobicity alone guarantees trichome retention.

Why "cannabinoids are hydrophobic, so gentle washing does not remove the active resin" is not a valid conclusion: "hydrophobic → won't dissolve" is true and well-supported. "Won't dissolve → washing doesn't remove the resin" is a non sequitur, because dissolution and physical/mechanical detachment are two different mechanisms, and the best available cannabis-specific evidence on the mechanical side (trimmethod trichome damage, ice-water hash mechanics) shows detachment under exactly the conditions — cold water, agitation — the claim is being used to dismiss as safe.

What remains genuinely, explicitly unknown and unresolved until studied directly on cannabis: microbial/fungal reduction from washing on real harvested flower; trichome retention/damage from washing, before and after subsequent drying; cannabinoid retention; terpene retention; post-wash drying behavior and re-wetting/regrowth risk, given that moisture control is this paper's own single strongest post-harvest risk lever (Section 6); and cross-contamination risk in shared wash water, by analogy to the produce-washing evidence in Section 9. Anecdotal grower practice (peroxide/lemon-juice/baking-soda bucket protocols) is widespread and internally inconsistent, with no cited scientific backing on either side — reported here as background context on how growers currently think about washing, not as evidence.

12. The World the Plant Grows In Has Changed — But Not in Every Way

The question: does credible evidence support the idea that cannabis today encounters environmental exposures that were absent, less prevalent, or less recognized in previous decades — and where it does, is that exposure surface-level (post-harvest-cleaning-relevant), internal/systemic (not addressable by washing), or simply unconnected to cannabis by direct evidence?

Only the findings that survived independent audit are retained here. The evidence does not support a blanket claim that "the environment is dirtier today." The best-measured national environmental trends actually run the opposite direction: EPA data show ambient PM2.5 down 46% and PM10 down 36% since 2000, and total sulfur and oxidized-nitrogen atmospheric deposition down 82% and 59% respectively in the eastern US since the early 2000s — real decreases from real regulatory controls, not detection artifacts. Any version of this section's argument that implies general environmental degradation across the board is not supported and should not be used.

Wildfire smoke and ash is the strongest, most defensible "things have changed" claim available, and it is the one surface-relevant exposure in this section with direct, cannabis-specific evidence. California's population-weighted wildfire-attributable PM2.5 rose roughly 8-to-9-fold between 2006 and 2018, a genuine, measured physical increase attributed to warming, drought, and fuel accumulation. This is the one issue in this section with a pre-existing, cannabis-specific, peer-reviewed citation already used elsewhere in this paper (Section 7): 57% of California cannabis growers report washing wildfire ash/particulate off growing plants as a field mitigation practice (Martin et al. 2025). This is the cleanest bridge available between "the environment changed" and "growers already respond by washing."

Modern agricultural pesticide toxicity profile — real change, general context, cannabis-specific gap. Total pounds of pesticide applied nationally is not a "more than ever" story — it peaked around 1981 and has declined somewhat since. What has genuinely changed is the toxicity profile: a peer-reviewed study found acute insecticide oral-toxicity loading on US agricultural land rose roughly 48-fold between 1992 and 2014, driven by a shift to neonicotinoids — lower volume, far higher toxicity and persistence per unit applied (DiBartolomeis et al. 2019). No study was found that documents or quantifies pesticide drift landing on cannabis flower specifically; this should be used only as general agricultural context, not as cannabisspecific evidence.

PFAS, heavy metals, and hop latent viroid (HLVd) are real, and in HLVd's case genuinely new, threats — but none of them are legitimate post-harvest washing claims, and this distinction must be maintained rigorously. PFAS compounds have been manufactured since the 1940s; what changed recently is regulatory recognition, not the chemicals' age. Hemp/cannabis is documented to take up PFAS from contaminated soil through its roots, including measured bioaccumulation in hemp pollen — every cannabis-specific PFAS source describes root uptake and internal accumulation, a systemic, cultivation-stage issue, never a surfacedeposition one. Cannabis's capacity to accumulate cadmium, chromium, and nickel from soil is wellestablished, decades-old science (Linger, Ostwald & Haensler 2002), also root uptake. Hop latent viroid (HLVd) was first reported causing disease in commercial cannabis in California in 2019 and has since become widespread in North American cannabis production (Bektaş et al. 2019; Adkar-Purushothama, Sano & Perreault 2023). HLVd transmits mechanically within plant tissue via cuttings, contaminated tools, and sapto-sap contact — a systemic infection replicating inside the plant's own cells. It is unambiguously a propagation- and tool-hygiene problem, not a surface contaminant, and must never be connected to a post-harvest wash product's mechanism of action.

Overall verdict

Rating: B — Good, not A — Strong. The proposition "the environment cannabis is grown in today is meaningfully different from decades ago in ways relevant to harvested flower cleanliness" is defensible only for a narrower set of claims than a blanket reading would suggest, and only wildfire smoke/ash lands as both genuinely changed and surface-relevant with direct cannabis-specific support. The grower conversation this section can responsibly support — "we didn't wash our buds back then; true, but are we growing in the same environment today?" — works only if it leads with wildfire smoke and is honest that most other "world has changed" candidates are either not surface-relevant (PFAS, heavy metals, HLVd) or not yet cannabis-specific (general pesticide toxicity shift, microplastics, industrial deposition).

13. Regulatory and Industry Recognition

Post-harvest microbial contamination of cannabis is formally, unambiguously recognized as a legitimate industry and regulatory concern — this is the single highest-confidence claim in the entire paper.

ASTM International's cannabis committee (D37) finalized ASTM D8575-25, a standard guide specifically for lowering microbial load on post-harvest cannabis and hemp inflorescence, in May 2025. AOAC International has published performance requirements specifically for yeast/mold enumeration on cannabis (SMPR 2021.009) and validated a standard method (AOAC 997.02) for the cannabis matrix, published in the Journal of AOAC International in 2023.

State regulatory practice is real but genuinely inconsistent, and no claim in this paper should imply uniformity across legal states. California requires inhalable cannabis products to be non-detected for four Aspergillus species and for Salmonella and STEC in a 1-gram sample. Michigan sets total-yeast-and-mold limits between 10,000 and 100,000 CFU/g depending on product category, with Aspergillus non-detect and a mycotoxin action limit triggered by an Aspergillus failure. New York similarly enforces Aspergillus zerotolerance with category-specific yeast/mold limits. Washington regulates bacteria only and has no yeast, mold, or Aspergillus testing requirement at all — a directly confirmed, useful illustration that state practice is inconsistent, not a unified national standard, and this paper corrects any prior framing that implied otherwise.

A 2025 peer-reviewed study is worth flagging as a live, contested scientific question rather than settled fact: it found that irradiated cannabis that had already passed standard CFU-based regulatory testing still carried detectable fungal DNA and mycotoxin residue, arguing current CFU-based testing may understate risk (Rani et al. 2025) — treated in full in Section 14. This should not be used to imply compliant product is unsafe; it is one research group's argument that the testing method, not the product, may need reconsideration, but it demonstrates post-harvest microbial control is an actively developing area of cannabis science, not a solved problem.

Rating: A (Strong) for the claim that post-harvest microbial contamination is formally recognized as a legitimate concern by regulators and standards bodies. This rating does not extend to any claim of regulatory consensus on specific numeric limits — state practice varies widely, and this paper does not claim otherwise anywhere.

14. Prevention vs. Remediation

This paper draws a clean, load-bearing distinction between two different strategies for arriving at compliant, clean flower:

Prevention: controlling conditions — moisture, temperature, airflow, handling hygiene, bioburden — before contamination becomes established. Prevention is a process cost, budgeted into growing the crop correctly the first time, regardless of outcome.

Remediation: treating finished or already-contaminated product after the problem exists — most commonly irradiation, ozone treatment, or non-thermal plasma. Remediation is a recovery cost, layered on top of the full cost of growing the crop, that only exists because prevention already failed.

Both paths can produce compliant flower. They are not the same path, and this paper does not treat them as interchangeable.

What ionizing-radiation remediation actually does. Gamma irradiation, electron-beam (e-beam) processing, and X-ray decontamination all work on the same underlying principle: ionizing energy damages microbial DNA (and other cellular structures), preventing replication and inactivating bacteria, mold, and fungal spores, without applying meaningful heat to the product. This is the same category of process the FDA has approved for produce, spices, and ground beef — with the important caveat that this is an approval of the technology for those specific foods; FDA has not evaluated or approved cannabis irradiation specifically, since cannabis remains outside FDA's federal regulatory purview. More than 60 countries currently permit food irradiation for one or more products (IAEA).

The peer-reviewed record on cannabis specifically is genuinely reassuring on the core cannabinoid-stability question. Hazekamp's controlled study across four pharmaceutical-grade cannabis varieties found no change in THC or CBD content after standard gamma treatment (minimum 10 kGy) — though the same study found real, substantial reductions in key monoterpenes (10–38%, cultivar-dependent), a finding that should always be cited alongside the cannabinoid-stability result, not detached from it (Hazekamp 2016). A 2023 study processing 2,016 kg of commercial biomass across 150 barrel lots reached the same cannabinoidstability conclusion at real commercial scale — total yeast/mold counts fell from 4,500–13,000 CFU/g to under 100 CFU/g, bacterial loads similarly collapsed, and E. coli/Salmonella were non-detect before and after — with no statistically significant change in terpene content (p=0.13) at a comparable or higher dose than Hazekamp's study (Majumdar et al. 2023).

These two best cannabis-specific studies disagree with each other on terpenes, and this paper reports that disagreement honestly rather than resolving it into a single "modest, evaporation-driven loss" narrative. Neither study screened for novel radiolytic degradation compounds, so "no new breakdown compounds were found" is an untested claim, not a confirmed negative result, in either study. No source anywhere in this paper's evidence base supports the specific comparison sometimes made that irradiation-driven terpene loss is "the same as flower sitting out on a table for about a week." That comparison is not traceable to any study and should not be used.

E-beam processing achieves comparable or better microbial log-reductions than gamma in a fraction of the time — roughly 15 minutes per batch versus roughly 24 hours for gamma — and does not require a radioactive source (Jerushalmi et al. 2020; Schnabel et al. 2026). Gamma-source systems carry a genuine structural cost X-ray-based systems avoid: the radioactive source cannot be switched off and continues decaying for the unit's life, requiring ongoing security and eventual disposal — an independently confirmed physical/regulatory fact, not a vendor-only claim (NCBI Bookshelf). The net economic conclusion from this — which technology is cheaper overall — is genuinely contested in the industry literature, with the gamma industry's own trade association publishing cost comparisons that favor gamma at commercial scale; this paper treats any specific gamma-vs-X-ray total-cost-of-ownership claim as a contested industry position, not a settled fact.

Cost is a real but secondary and less certain part of this argument. Retesting fees are commonly cited in the $200–500-per-panel range, but the best-sourced figure for this comes from a single consulting firm's unsourced estimate; a genuinely independent, peer-reviewed source (a 2020 study of California cannabis testing-compliance costs) confirmed that testing costs are highly sensitive to batch size and failure rates, and that destruction of failed product can exceed the direct cost of laboratory testing — supporting the general shape of the "true cost exceeds the sticker price" argument with better, if regionally and temporally limited, sourcing (Valdes-Donoso, Sumner & Goldstein 2020). Equipment throughput for dedicated cannabis irradiation systems — roughly one to fifty pounds per multi-hour cycle depending on model and dose — is directly confirmed against primary vendor specifications and corroborated by comparable competing equipment.

Labeling and disclosure requirements are real but jurisdiction-specific, not a uniform national rule. Utah and Nevada are independently confirmed by primary regulatory text to require irradiation-specific labeling. Connecticut joined that group in 2026: Public Act 26-8 authorizes remediation of cannabis flower or other cannabis plant material by ionizing radiation after microbial-test failure and requires packaging to disclose that the material underwent remediation by ionizing-radiation exposure. California does not require a comparable consumer-facing irradiation label. Maine's legislature considered and rejected a mandatorylabeling bill (LD 1567) covering irradiation and ozonation; it died in committee in 2025 after growers argued a radiation label would mislead consumers into thinking the product was unsafe. The cannabis trade press has documented growers using terms such as "cold pasteurized" and "electronically pasteurized" in discussing irradiation (MJBizDaily 2023). Most other legal-cannabis states' disclosure requirements were not independently confirmed in either direction during this paper's research and should be treated as an open question, not assumed to fall on either side.

Consumer psychology is real, general science — not a measured cannabis-consumer finding, and this paper does not present it as one. Peer-reviewed food-consumer research finds that the word "radiation"/"ionizing radiation" triggers dread associations independent of actual dose, and that a persistent minority of consumers mistakenly believe irradiated food becomes radioactive itself (Junaedi, McNeill & Hamlin 2024). This sits on top of a well-established, independently replicated "naturalness bias" in consumer food psychology: people rate minimally processed products as healthier, safer, and better-tasting largely independent of what the underlying science says (Meier, Dillard & Lappas 2019). Specific percentages exist in the literature — one study found only about 19% of respondents could correctly reject the "becomes radioactive" myth, with roughly half still willing to purchase once informed — but that specific study surveyed a region of Kazakhstan with an unusual, non-transferable historical relationship to nuclear exposure (Orynbekov et al. 2025), and no cannabis-specific consumer research on irradiation attitudes exists anywhere. The psychological mechanism (dread association, naturalness bias) is real and plausibly transferable to any product category; its specific magnitude and direction in cannabis consumers is genuinely unmeasured and this paper does not assert a number for it.

The key structural limitation, and the actual evidence-based case for prevention: irradiation reliably inactivates organisms, but it does not reliably undo what those organisms already did before treatment. A 2025 peer-reviewed study found that gamma-irradiated cannabis — including licensed-producer commercial product that passed standard CFU-based compliance testing — still carried detectable mycotoxins (aflatoxins, ochratoxin A, deoxynivalenol, T-2 toxin) and stable residual fungal biosynthetic gene signal (Rani et al. 2025). The mechanism is independently corroborated: toxins are small, stable metabolite/protein structures that are inherently harder to destroy with radiation than living cells are to inactivate; irradiation is designed to stop reproduction and kill organisms, not to chemically neutralize toxins already synthesized before treatment (Schnabel et al. 2026). A product can be irradiated, pass standard compliance testing, and still carry measurable toxin residue from contamination that occurred before the remediation step.

An Ounce of Prevention Beats a Pound of Radiation

This is not a claim that irradiation is ineffective or unsafe. Nothing in this paper's evidence supports that. Properly dosed gamma, e-beam, and X-ray decontamination reliably kill microbial contaminants without measurably damaging cannabinoids, and for some operations — particularly high-volume medical supply chains with strict pharmaceutical-grade requirements — it remains the most practical tool available.

The evidence-supported argument is narrower and more precise: remediation happens after contamination has already occurred, and some consequences of that contamination — pre-formed mycotoxins chief among them — can remain after treatment even when the treatment works exactly as designed and the product passes standard compliance testing. Prevention that keeps moisture, bioburden, and time-toestablishment low in the first place addresses the problem before that specific trade-off exists. An ounce of prevention, built into the drying and handling process from day one, avoids a category of problem that a pound of remediation — however well it works — cannot always fully undo.

15. Controlled Drying as Prevention

If contamination risk is concentrated in the drying and curing window (Section 6), the highest-leverage prevention intervention is controlling that window tightly — moisture, temperature, and airflow — rather than growing the crop the same way and treating the result afterward. This section discusses environmentally controlled drying and curing as a technology category, using vendor-neutral, independent evidence. It does not name, require, or imply endorsement of any specific commercial system.

The strongest available evidence is a 2024 peer-reviewed academic study conducted independently of any single commercial vendor (Israel's Agricultural Research Organization/Volcani Center and Hebrew University): controlled-atmosphere drying — managed temperature, humidity, and gas composition (CO₂/O₂/N₂ blends) — preserved terpene content significantly better than traditional open-air drying, with minimal, non-statistically-significant monoterpene/sesquiterpene loss under optimal controlled conditions versus greater losses under conventional drying. The same study found controlled-atmosphere drying better preserved cannabinoid concentration — conventional/open-air drying caused 3.5-fold higher THCA-to-THC degradation than controlled conditions — reduced total drying/curing time by at least 60%, and, most strikingly, eliminated visible mold entirely, while roughly 80% of inflorescences under traditional openair drying were infested with Alternaria alternata (Shimshoni, Birenboim et al. 2024). This is a genuinely strong, independent, category-level data point.

This is consistent with, and mechanistically explained by, evidence already established in Section 6: hot-air drying achieved up to a 2-log CFU/g reduction in total yeast/mold counts compared to slow ambient or freeze drying, from drying method alone, no decontamination step required (Baek, Grab & Chen 2025). The general scientific mechanism connecting tighter environmental control to better outcomes is water activity's wellestablished role in restricting mold growth and slowing cannabinoid degradation (Gwinn et al. 2023; FAO grain-storage guidance).

This category claim should be framed precisely: independent, peer-reviewed research demonstrates that precisely controlled drying environments — temperature, humidity/RH, and in the strongest study, atmospheric gas composition — outperform ambient/conventional room drying for both terpene and cannabinoid retention and for mold suppression, as a category-level finding, not as validation of any single commercial product's specific mechanism, target range, or marketing claims. No commercial controlled-drying system is required, named, or implied to be necessary by this paper.

Washing (Section 11) addresses a different part of the same problem — surface-level contaminants present on the plant before it enters the drying room — and, combined with controlled drying, forms a conceptual

"prevention stack": reduce incoming bioburden, then control the conditions that would otherwise let anything remaining establish itself. This paper's own evidence (Section 11) requires that this framing be stated carefully: the washing half of that stack remains a plausible, unverified-for-cannabis intervention, not a demonstrated one, and nothing in this section should be read to imply otherwise.

16. What the Evidence Does NOT Establish

Explicitly, to keep this document honest and to prevent overreach downstream:

• A generalizable, population-level percentage of commercial cannabis carrying non-visible fungal, microbial, or other contamination. No source supports a number at that scope.
• That detected fungal or bacterial DNA/sequence reads represent viable, infectious organisms. Detection and viability diverge substantially and repeatedly in the cited literature.
• That any specific mycotoxin or pathogen level documented in this paper has been linked to illness in an ordinary, healthy consumer under routine, regulation-compliant conditions. The only confirmed mass-illness event involved grossly elevated contamination from an unusual source, not routine handling.
• That indoor cultivation is cleaner than outdoor, or vice versa, as a general rule. The evidence supports "different," not "worse" or "better," in either direction.
• That washing harvested cannabis with water reduces microbial load, removes pesticide residue, preserves or damages trichomes, or affects cannabinoid or terpene content. No cannabis-specific study exists on any of these questions. This is the largest single evidence gap identified in this paper.
• That cannabis trichomes function as "flypaper," or that any direct evidence shows particulate capture by cannabis resin. This remains a plausible, mechanistically-grounded inference, not a proven finding.
• That "the environment cannabis grows in today is dirtier" as a blanket claim. The best-measured national environmental trends run in the opposite direction; only wildfire smoke/ash is both genuinely changed and cannabis-relevant on the surface-exposure question.
• That gentle washing preserves trichomes or potency. The best available cannabis-specific mechanical evidence points the other way, and no washing-specific study exists to confirm either outcome.
• That irradiation is ineffective or unsafe as a remediation category. The evidence is the opposite: properly dosed irradiation reliably reduces microbial load without measurably affecting cannabinoid content.
• That irradiation eliminates all consequences of prior contamination. It does not reliably eliminate preformed mycotoxins, even when it fully inactivates the organisms that produced them.
• That every legal cannabis state applies the same testing or labeling standards. State practice varies substantially — Washington tests no fungal organisms at all; Utah, Nevada, and Connecticut are confirmed here to require irradiation-specific product labeling or disclosure.
• Any claim, express or implied, about Terps USA Bud Wash's efficacy. This paper establishes the problem landscape and the state of remediation and prevention science only; it does not evaluate any specific commercial product.

17. Research Gaps and Future Directions

The largest evidence gap identified in this review is the absence of a controlled, cannabis-specific washing study. The following research questions would materially improve the evidence base, listed in order of scientific and practical importance:

1. Microbial/fungal reduction from washing, before vs. after, on real harvested cannabis flower — culture-based and PCR-based counts, using a defined wash protocol, compared against an unwashed control from the same harvest lot. (Highest research priority because it directly addresses the paper's largest unresolved evidence gap.) 2. Trichome retention/damage from washing — microscopy-based quantification of trichome density/integrity before and after washing and after subsequent drying, compared to unwashed control. (Directly tests whether washing affects trichome integrity, an unresolved question not answered by cannabinoid hydrophobicity alone.) 3. Post-wash drying behavior and re-wetting/regrowth risk — moisture/water-activity time-course after washing, and microbial re-growth monitoring if drying is delayed or incomplete, since this paper's strongest single finding (Section 6) is that moisture control is the dominant driver of postharvest fungal risk. (This is a critical safety and process question because any wet treatment changes the starting conditions for drying.) 4. Cannabinoid retention — potency testing on washed vs. unwashed flower from the same lot, postdry. 5. Terpene retention — terpene panel testing on washed vs. unwashed flower from the same lot, postdry. 6. Particulate/debris removal quantification — direct before/after measurement of visible and labdetected foreign material (dust, insect fragments, ash) on washed vs. unwashed flower. 7. Cross-contamination risk in shared wash water, by direct cannabis-specific measurement rather than produce-science inference, given the documented risk of disinfectant capacity being consumed across successive batches in the produce-washing literature. 8. A genuine, matched indoor-vs-outdoor comparative contamination study — same genotype, same season, simultaneous cultivation, to replace the current patchwork, single-facility evidence base.

18. Conclusion

Post-harvest cleanliness is a legitimate, evidence-backed problem category, not a manufactured one. A dedicated 2025 ASTM standard, validated AOAC testing methods, and state regulatory testing regimes all independently confirm that the industry and its regulators treat post-harvest microbial control as a real and unresolved area of concern.

Visual inspection alone is an insufficient basis for microbiological confidence. This is true in food safety generally and true specifically for cannabis, where fungal organisms can colonize plant tissue endophytically, with zero external symptoms, and where individual spores and cells sit one to two orders of magnitude below the resolution of the human eye. Regulators require laboratory testing instead of visual grading for exactly this reason.

Prevention continues after the plant is cut. Harvest and post-harvest handling — trimming, drying, curing, equipment hygiene — are independently, measurably implicated in microbial outcomes, separate from and in addition to cultivation-stage conditions. Drying speed and final moisture content stand out as the single most consequential, most controllable lever available to any grower, evidenced by real log-reductions in microbial load from drying method alone.

Washing is a plausible intervention category — strongly supported by analogy to produce science, particularly for structurally similar leafy greens and herbs — but cannabis-specific efficacy research remains genuinely incomplete. No study has washed harvested cannabis flower and measured the result. Every claim this paper makes about washing is built from cross-crop inference and cannabis chemistry, clearly labeled as such, not from direct cannabis evidence, and several common grower claims about washing's safety for trichomes and potency do not survive scrutiny.

Controlled drying and moisture management is one of the strongest demonstrated prevention levers available, independently supported by vendor-neutral, peer-reviewed academic research showing dramatic mold-suppression and quality-retention advantages over conventional ambient drying.

Remediation has a valid role. Irradiation works as designed, reliably reduces microbial load, and does not measurably damage cannabinoid content — none of this paper's evidence disputes that. But remediation is applied after contamination has already occurred, and a 2025 peer-reviewed finding shows that some consequences of that contamination — pre-formed mycotoxins specifically — can persist even through treatment that otherwise passes standard compliance testing. Prevention aims to avoid ever reaching that point. An ounce of prevention beats a pound of radiation not because the radiation doesn't work, but because prevention avoids a category of problem that radiation cannot always fully undo.

A reader who follows a citation in this paper should find a real, checkable source that supports the claim being made. That standard has been applied throughout this review.

19. References

References are listed alphabetically by first author or issuing body. Cross-crop and non-peer-reviewed sources are identified where relevant.

Adal AM, et al. (2025). "Cannabis sativa: origin and history, glandular trichome development, and cannabinoid biosynthesis." Horticulture Research 10(9):uhad150. https://doi.org/10.1093/hr/uhad150

Ahmed MZ, McKenzie CL, Osborne LS (2024). "Arthropod and mollusk pests of hemp, Cannabis sativa (Rosales: Cannabaceae), and their indoor management plan in Florida." Journal of Integrated Pest Management 15(1):1. https://doi.org/10.1093/jipm/pmad028

Al Ubeed HMS, Wills RBH, Chandrapala J (2022). "Post-Harvest Operations to Generate High-Quality Medicinal Cannabis Products: A Systematic Review." Molecules 27(5):1719. https://doi.org/10.3390/molecules27051719

ASTM International. ASTM D8575-25, "Standard Guide for Techniques to Lower Microbial Load of PostHarvest Inflorescence of Cannabis and Hemp" (2025). https://store.astm.org/d8575-25.html

AOAC International. SMPR 2021.009 (performance requirements for yeast/mold enumeration in cannabis). https://www.aoac.org/wp-content/uploads/2021/06/SMPR-2021_009.pdf ; AOAC Official Method 997.02 as validated for cannabis flower, Journal of AOAC International 106(2):412 (2023). https://doi.org/10.1093/jaoacint/qsac140

Baek Y, Grab H, Chen C (2025). "Postharvest Drying and Curing Affect Cannabinoid Contents and Microbial Levels in Industrial Hemp (Cannabis sativa L.)." Plants 14(3):414. PMID 39942976. https://pmc.ncbi.nlm.nih.gov/articles/PMC11821007/

British Columbia Ministry of Agriculture and Food. "Diseases of Cannabis in British Columbia" (fact sheet). https://www2.gov.bc.ca/assets/gov/farming-natural-resources-and-industry/agriculture-andseafood/animal-and-crops/plant-health/disease-factsheets/specialtycrops/diseases_of_cannabis_in_british_columbia.pdf

Adkar-Purushothama CR, Sano T, Perreault J-P (2023). "Hop Latent Viroid: A Hidden Threat to the Cannabis Industry." Viruses 15(3):681. PMID 36992390. https://doi.org/10.3390/v15030681

Bektaş A, Hardwick KM, Waterman K, Kristof J (2019). "Occurrence of hop latent viroid in Cannabis sativa with symptoms of cannabis stunting disease in California." Plant Disease 103:2699. https://doi.org/10.1094/PDIS03-19-0459-PDN

Benedict K, Thompson GR, Jackson BR (2020). "Cannabis Use and Fungal Infections in a Commercially Insured Population, United States, 2016." Emerging Infectious Diseases 26(6):1308–1310. PMID 32441624. https://wwwnc.cdc.gov/eid/article/26/6/19-1570_article

Britt K, Kuhar TP, et al. (2021). "Pest Management Needs and Limitations for Corn Earworm (Lepidoptera: Noctuidae), an Emergent Key Pest of Hemp in the United States." Journal of Integrated Pest Management 12(1):34.

Buirs L, Punja ZK (2025). "Endophytes in Cannabis sativa: Identifying and Characterizing Microbes with Beneficial and Detrimental Effects on Plant Health." Plants 14(8):1247. PMID 40284136. https://doi.org/10.3390/plants14081247

California Code of Regulations, Title 4, §15720 (Microbial Impurities Testing) and §15722 (Foreign Material Testing). https://law.cornell.edu/regulations/california/4-CCR-15720 ; https://www.law.cornell.edu/regulations/california/4-CCR-15722

CDC. "Salmonellosis traced to marijuana — Ohio, Michigan." MMWR (1981). PMID 6789127.

CDC. "Fatal Occupational Asthma in Cannabis Production — Massachusetts, 2022." MMWR 72(46) (2023). https://www.cdc.gov/mmwr/volumes/72/wr/mm7246a2.htm

CIDRAP (2025). "Arizona announces recall of marijuana products because of Aspergillus, Salmonella." https://www.cidrap.umn.edu/fungal-infection/arizona-announces-recall-marijuana-products-becauseaspergillus-salmonella

Comeau D, Novinscak A, Joly DL, Filion M (2020). "Spatio-Temporal and Cultivar-Dependent Variations in the Cannabis Microbiome." Frontiers in Microbiology 11:491. https://doi.org/10.3389/fmicb.2020.00491

Cornell Produce Safety Alliance — cleaning vs. sanitizing guidance. https://producesafetyalliance.cornell.edu/

Connecticut General Assembly (2026). Public Act 26-8 / Substitute H.B. 5350, "An Act Concerning Cannabis, Hemp and Infused Beverage Regulation." Primary legislative record and OLR bill analysis confirm authorization of ionizing-radiation remediation and a required disclosure label for remediated cannabis flower or plant material. https://www.cga.ct.gov/asp/CGABillStatus/cgabillstatus.asp?bill_num=HB5350&selBillType=Bill de Montagnac D, et al. (2026). "Reducing pesticide residues on produce: a scoping review of household produce washing methods." Frontiers in Environmental Health. https://doi.org/10.3389/fenvh.2026.1768399

DiBartolomeis M, Kegley S, Mineau P, Radford R, Klein K (2019). "An assessment of acute insecticide toxicity loading (AITL) of chemical pesticides used on agricultural land in the United States." PLOS ONE 14(8):e0220029. https://doi.org/10.1371/journal.pone.0220029

Falcon-Brindis A, Villanueva RT, Viloria Z, Bradley CL (2023). "A Novel Approach to Tally Aculops cannabicola (Acari: Eriophyidae) for Field and Laboratory Studies." Journal of Economic Entomology 116(2):591–598. https://doi.org/10.1093/jee/toad004

FAO. "Manual of the Prevention of Post-Harvest Grain Losses — Fungi and Mycotoxins" (chapter). https://www.fao.org/4/x5065e/x5065E0c.htm [general grain-storage science, cross-crop parallel]

FDA. Food Safety Modernization Act, Produce Safety Rule (water-quality standard). https://www.fda.gov/food/food-safety-modernization-act-fsma/fsma-final-rule-produce-safety ; FDA, "Buy, Store, and Serve Safe Food — Produce." https://www.fda.gov/food/buy-store-serve-safe-food/selecting-andserving-produce-safely ; FDA, "Food Irradiation: What You Need to Know." https://www.fda.gov/food/buystore-serve-safe-food/food-irradiation-what-you-need-know

Feder CL, Cohen O, Shapira A, Katzir I, Peer R, Guberman O, et al., Meiri D (2021). "Fertilization Following Pollination Predominantly Decreases Phytocannabinoids Accumulation and Alters the Accumulation of Terpenoids in Cannabis Inflorescences." Frontiers in Plant Science 12:753847. https://doi.org/10.3389/fpls.2021.753847

U.S. Government Accountability Office. "Meat Safety: Inspectors' Ability to Detect Harmful Bacteria Is Limited." GAO/T-RCED-94-228 (1994). https://www.govinfo.gov/content/pkg/GAOREPORTS-T-RCED-94228/html/GAOREPORTS-T-RCED-94-228.htm [food-safety parallel]

Gil MI, Selma MV, Suslow T, Jacxsens L, Uyttendaele M, Allende A (2015). "Effect of Disinfectants on Preventing the Cross-Contamination of Pathogens in Fresh Produce Washing Water." International Journal of Environmental Research and Public Health 12(8):8658–8677. https://doi.org/10.3390/ijerph120808658

Green BJ, Couch JR, Lemons AR, Burton NC, Victory KR, Nayak AP, Beezhold DH (2018). "Microbial Hazards During Harvesting and Processing at an Outdoor United States Cannabis Farm." Journal of Occupational and Environmental Hygiene 15(5):430–440. https://pmc.ncbi.nlm.nih.gov/articles/PMC6314012/

Gwinn KD, Leung MCK, Stephens AB, Punja ZK (2023). "Fungal and mycotoxin contaminants in cannabis and hemp flowers: implications for consumer health and directions for further research." Frontiers in Microbiology 14:1278189. https://doi.org/10.3389/fmicb.2023.1278189

Hazekamp A (2016). "Evaluating the Effects of Gamma-Irradiation for Decontamination of Medicinal Cannabis." Frontiers in Pharmacology 7:108. PMID 27199751. https://doi.org/10.3389/fphar.2016.00108

Holmes M, Vyas JM, Steinbach W, McPartland J (2015). "Microbiological Safety Testing of Cannabis." Cannabis Safety Institute white paper. https://cdn.technologynetworks.com/tn/Resources/pdf/microbiological-safety-testing-of-cannabis.pdf [nonpeer-reviewed industry white paper; used only for the visible-mold CFU threshold estimate]

Ikeda A, et al. "Role of Hydrophobins in Aspergillus fumigatus." PMC5872305. [general mycology, spore water-resistance mechanism]

International Atomic Energy Agency. "Food irradiation, benefits, use, standards." https://www.iaea.org/topics/food-irradiation

Jerushalmi S, et al. (2020). Gamma/e-beam dose-response study against Botrytis cinerea in cannabis. Journal of Cannabis Research. https://doi.org/10.1186/s42238-020-00020-6

Junaedi I, McNeill LS, Hamlin RP (2024). "Developing Food Consumer Attitudes towards Ionizing Radiation and Genetic Modification." Nutrients 16(20):3427. PMC11514585. https://doi.org/10.3390/nu16203427

Leung M, et al. (2025). "Evaluation of Fusarium Mycotoxins and Fungal Metabolites in Seized Cannabis in Arizona and California, 2023–2024." Environmental Health Perspectives 133(3-4). https://doi.org/10.1289/EHP16028

Lhotská M, Zemanová V, Pavlík M, Pavlíková D, Hnilička F, Popov M (2022). "Leaf fitness and stress response after the application of contaminated soil dust particulate matter." Scientific Reports 12:10046. https://doi.org/10.1038/s41598-022-13931-6 [lettuce — cross-crop parallel]

Linger P, Ostwald A, Haensler J (2002). "Heavy metal tolerance and accumulation of Cd, Cr and Ni by Cannabis sativa L." Plant and Soil. https://link.springer.com/article/10.1023/A:1026113905129

MacGowan C, Martynenko A (2025). "Electrostatic separator of cannabis trichomes: an innovative approach to extraction." Journal of Cannabis Research 7:23. PMID 40320547. https://doi.org/10.1186/s42238-02500281-z. [The paper describes conventional ice-water separation as using cold-water exposure and mechanical agitation to separate trichomes; its discussion of the detailed embrittlement mechanism relies partly on prior technical literature.]

Maine Legislature. LD 1567, 132nd Legislature, "An Act To Require Labeling Of Radiation Treatment And Ozonation Of Adult Use Cannabis And Inspection And Registration Of Associated Equipment" — died in committee, May 2025. https://fastdemocracy.com/bill-search/me/132/bills/MEB00015364/ ; opposition testimony record: https://legislature.maine.gov/testimony/resources/VLA20250505Esdaile133909362985484493.pdf

Majumdar et al. (2023). "Effect of Gamma Irradiation on Cannabinoid, Terpene, and Moisture Content of Cannabis Biomass." Molecules 28(23):7710. PMID 38067441. https://doi.org/10.3390/molecules28237710

Martin RW Jr, Wheeler K, Macdula CM, Jacobo J, Savage ER (2021). "Microbiological Implications of Commercial Cannabis in Northern California: A Survey of Bacterial and Fungal Contamination." Cannabis Science and Technology 4(7). https://www.cannabissciencetech.com/view/microbiological-implications-ofcommercial-cannabis-in-northern-california-a-survey-of-bacterial-and-fungal-contamination [trade publication, not peer-reviewed]

Martin JV, Dillis C, Starrs G, Schell D, Grantham TE, Butsic V (2025). "Wildfire impacts and mitigation strategies among California cannabis producers." PLOS ONE 20(4):e0321476. https://doi.org/10.1371/journal.pone.0321476

McKernan K, Spangler J, Helbert Y, Devitt-Lee A, Zhang L, Orphe W, Warner J, Foss T, Hudalla CJ, Silva M, Smith DR (2016). "Metagenomic analysis of medicinal Cannabis samples; pathogenic bacteria, toxigenic fungi, and beneficial microbes grow in culture-based yeast and mold tests." F1000Research 5:2471. https://doi.org/10.12688/f1000research.9662.1 ; McKernan K, et al. (2016). "Cannabis microbiome sequencing reveals several mycotoxic fungi native to dispensary grade Cannabis flowers." F1000Research 4:1422. https://doi.org/10.12688/f1000research.7507.2

McKernan K, Helbert Y, Kane LT, et al. (2021). "Whole genome sequencing of colonies derived from cannabis flowers and the impact of media selection on benchmarking total yeast and mold detection tools." F1000Research 10:624. PMID 34484691. https://pmc.ncbi.nlm.nih.gov/articles/PMC8383127/

McKernan K, McLaughlin S, Houde N, Kane LT, Helbert Y, Zhang L, Hom S (2022). "Next generation sequencing of cannabis flower samples and their respective quantitative PCR assays reveals false positive detection of off-target Aspergillus species." Zenodo. https://doi.org/10.5281/zenodo.7071829

Meier BP, Dillard AJ, Lappas CM (2019). "Naturally better? A review of the natural-is-better bias." Social and Personality Psychology Compass 13(8):e12494. https://doi.org/10.1111/spc3.12494

Michigan Cannabis Regulatory Agency. "Sampling and Testing Technical Guidance for Marijuana Products." https://www.michigan.gov/cra/-/media/Project/Websites/cra/bulletin/5Technical/Sampling_and_Testing_Technical_Guidance_for_Marijuana_Products_694124_7.pdf

MJBizDaily (2023). "Cannabis irradiation poses quandary for growers." https://mjbizdaily.com/news/cannabis-irradiation-poses-quandary-for-growers/370091/

MJBizDaily (2026). "Should cannabis remediated with radiation be labeled?" https://mjbizdaily.com/news/should-cannabis-remediated-with-radiation-be-labeled/615689/

Montville R, Chen Y, Schaffner DW (2001). "Glove barriers to bacterial cross-contamination between hands and food." Journal of Food Protection 64(6):845–849. PMID 11403136. https://doi.org/10.4315/0362-028x64.6.845 [food-safety parallel]

NCBI Bookshelf. Radiation Sources and Alternative Technologies in Medicine and Research (gamma-source decay, disposal, security cost basis). https://www.ncbi.nlm.nih.gov/books/NBK573887/ ; Toxicological Profile for Ionizing Radiation, "Mechanisms of Biological Effects." https://www.ncbi.nlm.nih.gov/books/NBK597563/

Nevada Cannabis Compliance Board. Regulation 12.065 (irradiation labeling). https://ccb.nv.gov/wpcontent/uploads/2021/05/Reg-12_rev2.pdf

New York Office of Cannabis Management. "Laboratory Testing Limits" (rev. Dec. 2025). https://cannabis.ny.gov/laboratory-testing-limits

NIOSH (2017). "Evaluation of Potential Hazards during Harvesting and Processing Cannabis at an Outdoor Organic Farm." Health Hazard Evaluation Report HHE 2015-0111-3271. https://doi.org/10.26616/NIOSHHHE201501113271

Olejar KJ, Park S-H (2022). "Industry-based misconceptions regarding cross-pollination of Cannabis spp." Frontiers in Plant Science 13:793264. https://doi.org/10.3389/fpls.2022.793264

Oregon State University Extension. "Aspergillus ecology and mitigation strategies in hemp production." https://extension.oregonstate.edu/community-vitality/health-outreach/aspergillus-ecology-mitigationstrategies-hemp-production

Orynbekov D, Kalibekkyzy Z, Nurgazezova A, Nurymkhan G, Kassenov A, Yermekov Y (2025). "Consumer Perception of Irradiated Food Products in the Abai Region of Kazakhstan." Foods 14(9):1625. PMC12071484. https://doi.org/10.3390/foods14091625

Pépin N, Punja ZK, Joly DL (2018). "Occurrence of Powdery Mildew Caused by Golovinomyces cichoracearum sensu lato on Cannabis sativa in Canada." Plant Disease 102(12):2644. https://doi.org/10.1094/PDIS-04-180586-PDN

Pulkoski M, Burrack H (2020). "Hemp Russet Mite." NC State Extension Publications. https://content.ces.ncsu.edu/hemp-russet-mite-in-industrial-hemp

Punja ZK, Collyer D, Scott C, Lung S, Holmes J, Sutton D (2019). "Pathogens and Molds Affecting Production and Quality of Cannabis sativa L." Frontiers in Plant Science 10:1120. https://doi.org/10.3389/fpls.2019.01120

Punja ZK, Ni L, Lung S, Buirs L (2023). "Total yeast and mold levels in high THC-containing cannabis (Cannabis sativa L.) inflorescences are influenced by genotype, environment, and pre- and post-harvest handling practices." Frontiers in Microbiology 14. https://doi.org/10.3389/fmicb.2023.1192035

Punja ZK, Sutton DB, Kim T (2023). "Glandular trichome development, morphology, and maturation are influenced by plant age and genotype in high THC-containing cannabis (Cannabis sativa L.) inflorescences." Journal of Cannabis Research 5:14. https://doi.org/10.1186/s42238-023-00178-9

Rani M, Kaddoura MJ, Samsatly J, Chamberland G, Jabaji S, George S (2025). "Detection of Mycotoxigenic Fungi and Residual Mycotoxins in Cannabis Buds Following Gamma Irradiation." Toxins 17(11):528. https://doi.org/10.3390/toxins17110528

Sapers GM, et al. (2006). "Efficacy of home washing methods in controlling surface microbial contamination on fresh produce." Journal of Food Protection 69(2):330–334. PMID 16496573. https://doi.org/10.4315/0362028x-69.2.330 [cross-crop parallel]

Schnabel U, et al. (2026). "Sanitation of medical cannabis flowers (Cannabis sativa L., flos): comparing current technologies and non-thermal plasma novelties from safety and quality perspectives." Journal of Cannabis Research 8(1):50. https://doi.org/10.1186/s42238-026-00417-9

Valdes-Donoso P, Sumner DA, Goldstein R (2020). "Costs of cannabis testing compliance: Assessing mandatory testing in the California cannabis market." PLOS ONE 15(4):e0232041. PMID 32324781. https://doi.org/10.1371/journal.pone.0232041

Shimshoni JA, Birenboim M, et al. (2024). "In Pursuit of Optimal Quality: Cultivar-Specific Drying Approaches for Medicinal Cannabis." Plants 13(7):1049. https://www.mdpi.com/2223-7747/13/7/1049

Spampinato G, Candeliere F, Amaretti A, Raimondi S, et al. (2024). "A three-years survey of microbial contaminants in industrial hemp inflorescences from two Italian cultivation sites." Journal of Cannabis Research 6:31. PMID 39020444. https://doi.org/10.1186/s42238-024-00241-z

Spitzer-Rimon B, et al. (2019). "Architecture and Florogenesis in Female Cannabis sativa Plants." Frontiers in Plant Science 10:350. https://doi.org/10.3389/fpls.2019.00350

Subramanya SH, Pai V, Bairy I, Nayak N, Gokhale S, Sathian B (2018). "Potassium permanganate cleansing is an effective sanitary method for the reduction of bacterial bioload on raw Coriandrum sativum." BMC Research Notes 11(1):124. https://pmc.ncbi.nlm.nih.gov/articles/PMC5812110/ [herb parallel]

Taylor DN, Wachsmuth IK, Shangkuan YH, et al. (1982). "Salmonellosis Associated with Marijuana: A Multistate Outbreak Traced by Plasmid Fingerprinting." New England Journal of Medicine 306(21):1249–1253. PMID 7070444. https://doi.org/10.1056/NEJM198205273062101

USDA Food Safety and Inspection Service. "Washing Food: Does it Promote Food Safety?" https://www.fsis.usda.gov/food-safety/safe-food-handling-and-preparation/food-safety-basics/washingfood-does-it-promote-food

Utah Administrative Code R66-2-14 (irradiation labeling). https://www.law.cornell.edu/regulations/utah/Utah-Admin-Code-R66-2-14

Washington Administrative Code 314-55-102. https://app.leg.wa.gov/wac/default.aspx?cite=314-55-102

WHO Expert Committee on Drug Dependence. Critical Review — THC water solubility (citing Yalkowsky et al.).

Zhao IY, Jung J, Moyne A-L, Schaffner DW, Harris LJ (2021). "Evaluation of glove type on survival and transfer of Escherichia coli in model systems and during hand harvesting of lettuce." JSFA Reports 1(1):17–25. PMID 35958100. https://doi.org/10.1002/jsf2.21 [produce parallel]