Short answer. Don't flush before harvest. It doesn't clean the bud, and it does harm. The last weeks of flower are when the plant packs on most of its final weight and finishes its trichomes and resin. Cutting its food then starves it during the most important phase, for no return. Controlled studies found that flushing did not improve flower quality, minerals or cannabinoids [1][2][3]. Longer nutrient withdrawal tended to cut dry mass and cannabinoid yield [2][3][4]. In the one taste trial, tasters preferred the unflushed flower [7]. Feed through to harvest at the strength the stage calls for. Taste and ash come from a healthy plant that was never overfed, followed by slow drying and proper curing.
House position: Myth. Pre-harvest flushing for taste or to "clean the plant" is a myth, and it's actively harmful: it starves the plant while it is building yield and resin. The only valid flush treats the medium, not the plant (see When flushing IS useful).
Greenleaf aside: flushing is for toilets.
What's happening: cannabis is a bioaccumulator
A bioaccumulator is a plant that takes up minerals and holds on to them. Cannabis is a well-known one (that's why hemp is used to clean contaminated land). Nutrients it absorbs don't sit in the bud like dust on a shelf, waiting to be rinsed off. They are built into tissue: proteins, nucleic acids, chlorophyll, cell walls, and ions held inside cells. Giving the roots plain water for a week or two doesn't pull those back out of the flower.
What plain water does do is force the plant to move mobile nutrients (N, P, K, Mg) out of its leaves to keep the flowers going. That's why fan leaves yellow fast during a flush. Growers see the yellowing and think "it worked". It didn't: in the Utah State work, leaf nitrogen fell but flower nitrogen didn't change [2]. The leaves were stripped; the bud was not cleaned.
Meanwhile, the plant is short of food during the weeks when bracts are swelling and trichome heads are filling. That's the harm.
Falling nitrogen isn't "cleaning", and some elements go up
When tissue nitrogen does fall in late flower, that reflects how the plant uses and moves N (remobilising it from leaves into developing flowers, and diluting it as flowers bulk up), plus the roots having fewer ions to take up. It isn't water washing the plant clean.
And flushing doesn't lower everything. Some elements rise in flower tissue after a flush, which is the opposite of what a "clean-out" would do:
- In the Clemson and Cannabis Research Coalition trials, calcium rose in flower with flushes of 3 weeks or more, and those same flushes cut dried bud weight. The authors note that tissue calcium commonly rises as nitrogen falls [9].
- In the 0–14-day industry trial, iron was at least 50 ppm higher in flower flushed 10 or 14 days, and zinc roughly 73 ppm higher at 14 days, while P, K, Ca, Mg and S didn't change and N was 6.7 % lower [10].
The likely reasons are a concentration effect (less bud mass with the same or slightly more of an element gives a higher percentage) and micronutrients released from the medium when plain water goes through it [9][10].being reviewed
The studies
| Study | Design | Result | Evidence quality |
|---|---|---|---|
| Stemeroff, Guelph 2017 [1] | Plain water for the last ≤ 2 weeks | No change in flower minerals or cannabinoids; no consistent yield change | MSc thesis, controlled |
| Hershkowitz, Utah State 2024 [2] | Pre-harvest nutrient deprivation | No gain in flower quality or total cannabinoid yield; deficiency symptoms within 1 week; leaf N fell, flower N unchanged; yield declined as flush length increased (as cited in [3]) | MS thesis, controlled (Bugbee lab) |
| Alden & Faust 2025 [3]; Alden 2025 [4] | 0 up to 3–4 weeks of fertiliser restriction | Flushing "tended to decrease dry mass, cannabinoid yield, and mineral nutrient concentrations" [4] | Peer-reviewed paper and PhD dissertation |
| Saloner et al. 2024 (as cited in [3]) | 2-week flush | No effect on inflorescence yield | Peer-reviewedbeing reviewed |
| Industry trial [7][10] | 0, 7, 10, 14 days | Most minerals unchanged; N 6.7 % lower, Fe and Zn higher after 10–14 days; tasters preferred unflushed | Single, industry-funded |
The pattern: short flushes do nothing useful; longer ones cost you. No controlled study has shown a gain in quality, cannabinoids or yield. As the HortScience authors put it, no study to date has shown that flushing improves smoking quality [3].
Late-flower nutrition matters
Cannabis yield responds to nutrient supply through flower. In Saloner and Bernstein's work, inflorescence yield and plant function rose with nitrogen supply up to about 160 mg/L N in flower [5]. You don't need luxury levels either: work from the Bugbee lab found that extra phosphorus and excess fertiliser did not raise yield or quality and just ended up in the runoff [2][6]. The target is "enough, all the way to the end", not "too much, then nothing".
The nitrogen and potency trade-off
Nitrogen also affects the percentage of cannabinoids. In the same work, cannabinoid and terpenoid concentrations were suppressed at higher N supply, while yield rose with N up to ~160 mg/L [5]. So lower N can raise cannabinoid and terpene percentages while lowering total yield: a smaller crop with a slightly higher number on the label. It's a trade-off you set with your feed recipe through flower (see nitrogen and feed by stage), not a reason to flush. Starving the plant at the end costs yield and resin without that being a controlled choice.
What actually drives taste and ash
- Plant health through the whole grow. A plant that was never hungry, never burnt, and never stalled.
- Not overfeeding in the first place. Sensible EC and P in late flower (see feed by stage). Excess salt is a root-zone problem, fixed with a root-zone flush when it happens, not with a pre-harvest starve.
- Slow drying: about 7–14 days in the dark at 15–21 °C (60–70 °F) and 55–62 % RH (see drying).
- Proper curing: sealed jars at 58–65 % RH, burped, for 2–8 weeks (see curing and burping).
Black ash and harsh smoke almost always trace back to fast drying, wet or rushed curing, or how the joint burns, not to nutrients left in the plant.
Instead of flushing: taper nitrogen
The GenetixBay approach is taper nitrogen, don't flush. In late flower, cut N back, towards zero near the finish, while keeping P, K, Ca, Mg, S and the micronutrients fed at normal strength. You get the controlled side of the N trade-off (lower N, a slightly higher cannabinoid and terpene percentage [5]) without starving the plant of everything else while it finishes its resin. Experienced growers widely recommend it; it's Grower practice, and no controlled trial has tested this exact ramp.
| Weeks before harvest | N (mg/L) | Everything else |
|---|---|---|
| −4 | 100–150 (normal late flower) | Full late-flower feed |
| −3 | ~75–110 (about 75 %) | P, K, Ca, Mg, S and micros unchanged |
| −2 | ~50–75 (about 50 %) | Unchanged |
| −1 | ~25–40 (about 25 %) | Unchanged |
| Final days | 0–20 | Unchanged; keep normal runoff |
A bottled 2-part can't do this: dilute it to cut N and you cut every other element with it. Single salts make this possible: drop calcium nitrate and potassium nitrate, take K from potassium sulfate and MKP, and switch the Ca source to an N-free salt such as calcium chloride (small doses only: 50 mg/L Cl brings just ~28 mg/L Ca, so keep Cl under ~50–70 mg/L; see chloride) or calcium sulfate (gypsum) dissolved directly in the final solution. A per-element dosing controller does the same job automatically. Details in nitrogen and feed by stage.
What to do in the final weeks
- Feed through to harvest at the strength the stage calls for, tapering only N (above). In coco, rockwool and hydro keep normal late-flower EC and normal runoff. In soil, keep your normal feeding schedule.
- Watch runoff EC. If it climbs well above input, that's salt build-up. Do a root-zone flush (below) and go back to normal feeding.
- Let the plant fade on its own. Some lower-leaf yellowing in the last 2–3 weeks is normal senescence. Don't force it.
- Judge harvest by trichomes (see trichome ripeness), then put your effort into drying and curing.
- Living soil: there's nothing to flush. Keep watering normally.
When flushing IS useful
There's one valid flush, and it treats the medium, not the plant. A root-zone flush washes accumulated salts out of the medium so you can start fresh. It can be done at any stage when there's a real problem. It isn't a pre-harvest ritual.
When to do it
- Runoff EC far above input (roughly input + 1.0 mS/cm or more in coco/soil; slab EC above input + 1.5 in rockwool) (see runoff testing).
- Salt crust on the medium surface, burnt tips spreading on fresh growth (see brown or burnt tips).
- Lockout: deficiency symptoms despite a full feed, with runoff pH or EC out of range (see pH and nutrient availability).
How to do it
- Measure first. Record input and runoff pH and EC so you can see the change.
- Choose the water: - Soil: pH-adjusted plain water (pH ~6.3–6.8). - Coco, rockwool, hydro media: a light nutrient solution (about EC 0.4–0.8 mS/cm, with Ca and Mg) at pH ~5.8–6.0. Plain water in coco strips Ca and Mg from the exchange sites and can trigger the deficiency you were trying to fix.
- Volume: run about 2–3 times the container volume through slowly, in several passes, letting it drain between them. For an 11 L (3 gal) pot, that's ~20–30 L.
- Check runoff EC on the last pass. Stop when it's close to your flush input (within ~0.3–0.5 mS/cm) and runoff pH is back in range.
- Resume normal feeding at the next watering (soil: once the top few cm dry back). Don't leave the plant on plain water.
- Find the cause: feed EC too high, too little runoff, or water too hard (see water quality).
Common mistakes
-
Stacking one myth on another: a flush followed by 24–72 h of darkness. Both cut the plant's resources in its final days.
-
Starving plants for 1–3 weeks before harvest and losing bulk and resin.
- Reading yellow fan leaves as proof the flush "worked".
- Blaming black ash on nutrients when the bud was dried in 3 days.
- Flushing coco with plain water and causing Ca/Mg deficiency.
- Drowning a soil pot that's already waterlogged.
- Using a root-zone flush as a substitute for fixing feed EC or runoff volume.
Greenleaf's standing challenge: darkness before harvest. Putting plants in 24–72 h of darkness before the chop is a Myth, and the founder's verdict is "complete and utter nonsense": why cut photosynthesis in the most critical part of the life cycle, while the plant is still finishing its resin? For more than 20 years the founder has offered a standing challenge: find a peer-reviewed study showing a benefit (not a single-plant study, not biased or skewed) and a 50-pack of seeds is yours, no questions asked. Nobody has claimed it.
Evidence
- Pre-harvest flushing improves taste, ash or "cleanliness": Myth (house position). No controlled study has shown a quality benefit [1][2][3], and the one sensory trial favoured unflushed flower [7].
- Flushing does not raise cannabinoids or yield: Supported [1][2][3][4].
- Longer nutrient withdrawal costs dry mass and cannabinoid yield: Supported [2][3][4]. Short flushes (≤ 2 weeks) didn't consistently change yield in [1] or Saloner et al. (via [3]), so the size of the loss depends on how long and how hard you starve the plant.
- Root-zone flushing to clear salt build-up: Grower practice, consistent with substrate-management guidance.
- Some flower elements (Ca, Fe, Zn) rising after a flush, a concentration effect: Supported by trial reports [9][10], not yet by a journal paper.
- Lower N raising cannabinoid and terpene percentages while lowering yield: Supported [5]. A trade-off, not a reason to flush.
