Harvest guide

When to Harvest Cannabis: Read the Trichomes, Then Decide

Cannabis harvest timing is best judged by reading trichomes, pistils, and canopy maturity together. Clear trichomes are early, cloudy trichomes often line up with the main harvest window, and amber trichomes signal a later finish. If you want the shortest definition first, start with what are trichomes and then come back here for the full harvest call.

This is one of those moments where the plant is unusually honest. The hard part is that growers love to interrupt that honesty with wishful thinking.

Clear Usually too early if the goal is mature, fully expressed flower.
Cloudy Often treated as the main harvest window, though it is not a proven potency peak.
Amber Later maturity. Growers often call it a heavier finish, but no study links amber share to any effect.
Visual reference

The flower can look ready before the trichomes fully agree, which is why the image is a cue and not the verdict.

Mature structure, curled pistils, and a more settled surface can all suggest that harvest is close. They help the grower know when to inspect seriously, but they do not replace magnification and a clean read across the canopy.

  • Use the overall look of the flower to know when inspection matters most.
  • Let the trichome balance decide whether the plant is early, in the window, or clearly later.
  • Keep the final call tied to what the buds show under magnification, not to how photogenic the cola has become.
Mature dried cannabis cola with frosty trichomes and curled orange pistils resting beside a brass jeweler's loupe on a cream backdrop.
Use the overall maturity of the flower as a cue, then confirm the actual harvest window with trichome inspection under neutral light and decent magnification. The words printed on this image (observed maturity, peak potency, harvest ready) are illustration text, not a ColaXpress observation or measurement, and trichome color has not been shown to pinpoint peak potency.
The short answer

When to harvest cannabis: inside a window, not on a set day

The plain answer to when to harvest cannabis: inside a window, not on a set day. By grower convention, not any tested rule, the window sits around the point where most trichome heads on the bracts are milky (cloudy) and the first amber shows, most pistils have darkened, and the calendar is near the breeder's flowering time. Trichomes tell you the plant is in that window; they don't pick the day. In the only study that tested trichome colour as a harvest-day predictor, it missed the lab-measured cannabinoid peak by more than guessing one average date for every plant.

An honesty note: ColaXpress has no grow of its own and counted no trichomes. Everything below comes from published studies, public grow logs and third-party images, all credited. The numbers are theirs; the reading is ours.

How to check

How to Check Trichomes Before Harvest

  • Start near the breeder's earliest date. Begin at the low end of the listed flowering time, counted from the flip to 12/12 (autoflowers: the listed seed-to-harvest days, from germination). Ran 12/12 from seed? Count from first flowers, which that page's documented grows put three to five weeks from seed.
  • Know the colours. Cloudy vs amber trichomes is the visual reference for each colour; the trichome visual guide shows heads up close.
  • Read the bracts. Look at heads on the calyxes (bracts), the pods that make up buds, not the sugar leaves. A 2025 review by Alberti and colleagues (Plants) names bracts as the most consistent site for trichome-density counts; no study shows sugar-leaf heads amber first.
  • Sample buds at different heights. Check the top cola, a mid-canopy bud and a low one. Heads on one bract mature out of step, and a grower below logged amber at the top about two weeks before the lower buds.
  • Check every 2–3 days from the earliest date (practice, not a tested interval).
  • Use the magnifier you have. Practitioner picks run from a 10x loupe (Cervantes) to a 30–60x loupe or USB microscope (a Cannabis Training University guide); no study has compared powers. A phone with an add-on 10x macro lens was the researchers' tool below, though they blamed phone images partly for weak trichome results; take several shots per bud. The trichome magnification guide compares tools.
  • Decide on a window, then a day. When most bract heads are milky, the first amber shows and most pistils have turned, you're in the conventional window. No study marks where it closes; in the Czech series below, a week past the peak cut THCA (THC's raw acid form) 12–22%. Even the best predictors below missed by about four days, so picking a day inside it to suit your schedule is reasonable.
  • Write it down. Date, flower day, which bud, a rough clear/milky/amber read and pistil state go in your grow logbook. Next run, you'll know this cultivar's window.
Phone macro images of cannabis flower surfaces with detected trichome heads boxed, beside single heads labelled clear, cloudy and amber by the study authors.
Lorberboym and colleagues (2026) shot these through a phone with a 10x macro lens: detected heads boxed, plus single heads the authors labelled clear, cloudy and amber. Colours here are not calibrated. Figure: Lorberboym, Lazare, Golshmid & Shani (2026), Agriculture 16(4):460, Fig. 6, CC BY 4.0. Resized.
Illustrated cannabis plant labeling where to check trichomes — mid-bud, calyx-rich surfaces, and several canopy samples — plus a clear-milky-amber legend.
Inspection is more accurate when it stays on calyx-rich flower surfaces, mid-bud sites, and multiple canopy samples instead of relying on sugar leaves alone. The chart's peak-potency title, milky-as-peak legend, 30x to 100x range and sugar-leaf warning are not tested findings: milky as peak is unproven, no study has tested loupe power, and no study has compared trichome color on sugar leaves and buds.
Trichome stages

The harvest window gets much easier to read when the stages stay visual.

Four trichome macros in a row progressing from mostly clear to mostly cloudy milky, then cloudy with some amber, and heavy amber heads.
Use this chart to train your eye. Mostly clear usually means wait, mostly cloudy often marks the active window, cloudy with some amber leans later, and heavy amber is clearly later.
The evidence

Why Trichomes Give You a Window, Not a Day

Lorberboym and colleagues (2026, Agriculture) photographed 30 flowers from six greenhouse cultivars through iPhones with a 10x macro lens (about 20–30x effective), had software sort the heads and lab-tested cannabinoids from day 51 to 79 after the flip. Milky ratios tended to peak near the cannabinoid maximum, and the authors read amber as mostly post-peak. Each cue's typical miss as a harvest-day predictor, in days (RMSE) for total cannabinoids and THC:

  • Average-date guess (day 65 after the flip, their mean sampling date, for every flower): 3.6 and 4.0.
  • Orange-pistil threshold: 4.5 and 4.4.
  • Milky-trichome threshold: 7.8 and 7.9.
  • Amber-trichome threshold: 8.7 and 7.4. The best amber cut-off for THC was just 5% of detected heads.

Cultivars differed: in one, THC peaked early and fell while amber kept climbing. Thirty flowers is thin evidence, but it's the only test of its kind.

Color to window

Turn what you see into a cut-now, wait, or later decision.

Flowchart branching trichome color into three calls: mostly clear means wait, mostly cloudy is the window many growers treat as peak, and more amber is a later harvest.
Trichome color to harvest window, as growers commonly read it: clear = wait, cloudy = the window many growers treat as peak, amber = later. The chart's potency and effect labels are grower lore: no study links amber heads to couch-lock or sedation, and in the one small study that scored color as a harvest-day predictor, simply using that study's average calendar day came closer to the lab-measured peak.
Save this one

The window at a glance.

The same three-stage read in a single vertical card you can keep on your phone for the inspection: mostly clear means too early, mostly cloudy is the harvest window, and amber-dominant means you have likely gone past the window. Growers often describe that finish as heavier or more sedative, but no study links the share of amber heads to any effect. Use it as a quick reference, then confirm under the loupe across several buds.

Vertical three-stage trichome macro card: too-early clear heads on top, the mostly cloudy harvest window in the middle, and amber-dominant heads below.
Too early, harvest window, too late, top to bottom.
Decision guide

Most harvest calls come down to cut now, wait, or recheck in a couple of days.

What you are seeing What it usually means Best next move
Mostly clear trichomes The plant is still early even if it looks tempting from across the room. Wait and recheck in a few days.
Mostly cloudy with little amber The main harvest window many growers aim for is opening. Cut now, or watch closely if you prefer a slightly later cut.
Cloudy with a noticeable amber mix The window has moved later. Growers often describe this finish as heavier, though no study links amber share to any effect. Cut if a later harvest is what you were aiming for.

Still the right move to wait when:

  • Most of the trichomes are still clear even though the plant looks visually tempting from a distance.
  • Pistils have darkened, but the buds themselves are still reading earlier than the room wants to admit.
  • The breeder timeline says one thing, but the plant under the loupe is still telling a slower story.
  • You are mostly feeling tired of waiting rather than convinced by what the flower is showing.

Harvest regret is usually born in that last category. If the evidence is mixed and the plant is close, a short wait and a cleaner recheck beat a rushed cut almost every time.

What the research says

The Amber Rules of Thumb, and Who Said Them

SourceHarvest ruleBasis
Rosenthal & Downs, Marijuana HarvestFirst amber heads, for most potencyPractitioner book
Cannabis Training University guide on edrosenthal.com (2025)80–90% milky, 5–15% amberTraining-school guide, no citations
Cervantes, Cannabis Encyclopedia (2015)Indoors, very few coloured heads; one caption cites about 10% amberPractitioner book
NC State Extension (CBD hemp)20–30% amber"Our best guess from observations last year"
Rosenthal, The Cannabis Grower's Handbook (2021; also in Marijuana Harvest)30/70 rule: about 30% amber, 70% milkyCalled traditional, possibly inaccurate
Lee and colleagues (2026), one CBDA cultivarMilky-dominant, under 30% amberCBDA peaked earlier; the window held 79–86% of it

They span first amber to about a third; none has been validated against lab potency.

What Early or Late Actually Costs

The best weekly drug-type data: a Czech group cut three plants a week of one cultivar ('McLove') under HPS. Against each nutrient treatment's THCA peak (Malíková and colleagues, 2024, Scientific Reports):

  • 1 week early: about 4% to 19% less THCA.
  • 2 weeks early: about 16% to 25% less.
  • 1 week late: about 12% to 22% less.

One cultivar, three plants per point: the sizes are illustrative; the shape is the lesson. Concentration rises, peaks, then plateaus or dips; in Aizpurua-Olaizola's 2016 indoor series, THC-type THCA peaked in flower week 9, then ran about 3% and 8% lower over the next two weeks (chart-read). Dry flower weight kept climbing: 7.7 to 25.1 g per plant from week 5 to 11 in a CBD-type trial by Crispim Massuela and colleagues (2022). In Tran and colleagues' 2025 study, cannabinoid yield per inflorescence (concentration × weight) peaked at the latest pistil stage in 18 of 25 strains. Early costs potency and weight still to come; late costs potency too, though weight often keeps rising.

Amber Is Gland Ageing, Not CBN

Clear to milky to amber tracks the gland maturing, then ageing. Punja, Sutton and Kim (2023, Journal of Cannabis Research) saw browning arrive as heads age, collapse and detach. What turns a head amber is unknown: oxidised phenolics (Punja) and a non-cannabinoid compound such as a flavonoid (Sutton and colleagues) are untested ideas.

Research figure with graphs of trichome head translucency falling and a redness score rising over flowering weeks, a macro of bracts with reddish-brown heads, heads circled by colour class, and visible versus UV views of a bract.
Punja and colleagues (2023): measured head translucency fell and a redness score rose from week 3 to week 7; week-8 bracts show reddish-brown heads; in one image, heads are sorted clear, milky or red-brown. Figure: Punja, Sutton & Kim (2023), Journal of Cannabis Research 5:12, Fig. 13, CC BY 4.0.

The classic gland-level data run against "cloudy = peak". Turner, Hemphill and Mahlberg (1978, UNODC Bulletin on Narcotics) measured THC per gland at 57 ng in clear mature glands, 35 ng in yellow aged ones and 9 ng in red-brown senescent ones, with CBN at none, 21 and 1 ng; their scheme has no cloudy stage. It's one 1970s analysis of a seeded greenhouse clone, and per-gland content isn't flower potency: flowers keep adding glands, so it's no case for cutting at clear. Its 21 ng of CBN in aged yellow glands is the one gland-level hint of CBN on the plant.

Modern whole-flower tests don't bear that out. Late harvest doesn't build CBN: Tran's team flash-froze 100 samples and found 0.09 mg/g or less at every pistil stage through fully amber pistils (trichome colour wasn't scored); Aizpurua-Olaizola detected none through flower week 11. A one-cultivar greenhouse study found no quantifiable CBN at harvest even where 36–47% of heads were dark brown (Stoilkovska Gjorgievska and colleagues, 2023). A 2019 outlier (Zager and colleagues, Plant Physiology) reported 0.3–1.7% CBN in fresh pre-amber buds, unexplained and unreplicated. CBN forms mainly in storage, by air oxidation: in Ross and ElSohly's room-temperature series, CBN as a share of THC rose from 2.5% after one year to 14.2% after four.

The sleep link is weak and mixed. In two trials, CBN missed its main sleep outcome: 30 and 300 mg in a 20-person sleep-lab study (Lavender and colleagues, 2026) and 20 mg in an industry-run trial of 293 (Bonn-Miller and colleagues, 2024), with secondary gains at most. No study links percent-amber flower to sedation or couch-lock.

Pistils: The Second Opinion

Pistils are stigma hairs; their "amber" isn't trichome amber, but it carries signal. Tran's team staged 25 strains by pistil colour alone; total cannabinoids peaked at the mostly-amber or fully-amber pistil stage in 22, with one flower per strain per stage. In Lorberboym's test, orange pistils beat both trichome thresholds and trailed the average-date guess by a day or less, beating it only for CBD (one cultivar).

Macro photo of sugar leaves on a cannabis bud densely covered in stalked trichomes, heads ranging from glassy to milky white, with rust-brown pistils.
Third-party macro of sugar leaves on a bud: stalked heads range from glassy to milky white, beside rust-brown pistils. Its source assigns no ripeness stage, and camera colour is not a calibrated reference. Photo: Cannabis Pictures, via Wikimedia Commons, CC BY 2.0. Resized.

Hemp extension triggers differ: University of Missouri Extension gives 70–80% orangish-brown; UVM Extension generally starts at about 90% browning, with other cues. Pistils can brown early: Punja's plants had reddish-brown stigmas by week 6 of 8.

What Growers Logged

Three public logs show reads drifting between plants, buds and breeder dates. Every number is the grower's own; GrowDiaries counts weeks from germination, and none states the magnification used.

  • Tygrow78, White Widow DWC (Growers Network forum). Closet tent, DIY recirculating DWC of about 12 gallons, three plants. They chopped on flower day 51 at about 5% amber on most buds; their most nutrient-deficient plant sat near 50/50 amber and cloudy (they don't blame the deficiency).
  • ApothecAries, Northern Hash Plant 2x2 Grow Tent 2023 (GrowDiaries). 2×2 tent, 100 W LED, soil, two plants. Their seed source said 7–8 weeks; on flower day 59 they logged no cloudy heads yet. They chopped at mostly cloudy.
  • plasticbeach, Mephisto Mango Smile (GrowDiaries). Autoflower, 100 W LED, soil. A few top buds began to amber in week 13; by week 14 the top sat near 10% amber while lower buds were still turning cloudy, and amber reached them in week 15. No harvest was posted.

Harvest-Day Myths: Darkness and Flushing

Handoff

The harvest call should already be thinking about drying and curing.

A cleaner harvest decision makes the next stages easier immediately. When the plant is cut at the right window, drying starts with fewer doubts and curing is less likely to become a long argument with the jar. Every later finish decision inherits this one. It is the call that sets the tone for everything after it.

Once the cut is made, the next question is no longer ripeness. It becomes moisture control. That is why the cleanest follow-up reads are how to tell buds are ready for jars and how to cure cannabis after drying. If the trichome balance is still your main question, the narrower explainer on milky versus amber heads is the right side path.

Once it's down, the next call is whether to trim wet or dry.

FAQ

FAQ

What percentage of amber trichomes should I harvest at?

No number is validated. Published rules run from first amber to about 30% amber, and in the only study that tested trichome-colour cut-offs as harvest-day predictors, an average-date guess beat them all. By convention, not proof, treat mostly milky heads with the first amber showing as the window.

Do amber trichomes mean more CBN or a sleepier effect?

Not in meaningful amounts. Careful analyses of freshly harvested flower find trace CBN at most, even with many dark-brown heads; it builds mainly in storage. No study links amber to sedation, and CBN's sleep evidence is weak and mixed.

Can I harvest by pistils instead of trichomes?

Not alone; read them alongside trichomes and the calendar. Pistil stage tracked the cannabinoid peak in most strains in one study and nearly matched an average-date guess in another, but pistils can brown early.

Does 24–72 hours of darkness before harvest help?

We found no study of it. Treat it as a ritual, not a tested technique.

Read the bracts, log the window, and let the calendar break the tie.