My leaves are curling. Is it heat or is it hunger?
Stop looking at the leaf. Look at where on the plant it is.
By Curtis Olson — retired software developer, home grower, and the builder of the device this site is about. More about me and why I built it. · Last reviewed
The short answer
Compare the top of the plant to the shaded leaves underneath.
If the leaves nearest the lamp are folding and the shaded ones lower down are flat and open, that is a light-and-heat problem. If the whole plant is curling top to bottom — including leaves that are nowhere near the lamp — then the light is not the explanation and the root zone is worth looking at.
The reason this works is simple enough to hold onto: roots feed the entire plant. A root-zone problem does not stop at the top layer. Radiant heat does, because it lands where the light lands.
Why the position test works
A curled leaf is not a symptom of one thing. Folding up along the middle — what growers call tacoing or canoeing — is what a leaf does to reduce the surface it is presenting. It is a response to load, not a request for food.
The load is heat the leaf cannot shed. And the leaf’s own temperature is not your air temperature, which is the part that makes this so hard to see: your thermometer can read perfectly normal while the leaf it is sitting next to is in trouble.
The best measurements we know of on that gap are Nelson & Bugbee (2015, PLOS ONE), who compared leaf and air temperature under sunlight, greenhouse light, HPS and LED:
- Plants that were not water stressed held their leaves within about 2 °C of air temperature under all four light sources.
- Swapping HPS for LED moved leaf temperature by only around 1.3 °C. The lamp type is a small term.
- The plant’s water status and its own evaporative cooling mattered far more than the light source did.
- Under water stress with high light and low airflow, leaves ran hotter than the air — by around 8 °C under LED, and more than that under HPS.
A leaf cools itself by sweating: water leaves through the pores underneath, evaporates, and carries heat away with it. A plant drinking freely sweats freely and sits about at air temperature. A plant that is short of water shuts those pores to conserve what it has, stops sweating, and then simply sits under the lamp getting hotter.
So the leaf-to-air gap is smallest when everything is fine and largest — and pointing the other way — exactly when it is not. That is why the curl shows up at the top first. Those leaves are taking the most light and have the least help.
The other test, which costs nothing and runs overnight
Look at the same leaves about two hours after the lights go off.
If they have relaxed flat in the dark, the load was coming from the lamp, and your levers are height and airflow. If they are still curled with the lights off and have been for hours, the lamp is not what is doing it, and you should be looking at water and roots after all.
It is not a perfect test — nothing done by eye is — but it is free, it runs while you sleep, and it separates the two most common answers.
| What you see | Top of the plant | Shaded leaves below | Two hours after lights-off |
|---|---|---|---|
| Light and heat | Folding | Flat and open | Relaxed |
| Water or root zone | Folding | Folding too | Still folded |
Droop is not claw, and the difference decides what you do
Two things get called the same name and they are not the same shape.
Droop is the stem the leaf hangs from. The whole leaf sags, but the blade itself is still flat and open. Shaded lower leaves on a plant that has been defoliated hang like this routinely. It is usually nothing.
Claw is the blade. The edges roll under, the tip hooks down, and the leaf goes tight. That one is worth chasing.
Mistaking the first for the second is a common way to end up treating a plant that is fine.
When only the top leaves taco, and the plant below is fine
This is the position test giving its clearest answer, and it is worth saying what it usually means in practice. The leaves closest to the lamp fold up along the midrib — the taco — while everything a hand’s width lower is flat. That is radiant load on the leaves that can see the lamp, and it happens at a fixed height and power that was fine a week ago because the plant grew into the light. The fix is distance or power, not feed; a leaf that has tacoed will not un-taco, so judge by the next leaves to open after you change it.
Two things make it worse at the same lamp: a hot patch of still air right under the fixture, and a canopy that has grown dense enough that the top leaves are the only ones moving air at all. Check both before you raise the light — a fan moved to sweep across the tops sometimes fixes what looks like a lamp problem.
Should I defoliate a plant that is canoeing?
Usually not yet, and the reason is the one this page keeps coming back to. A leaf canoes because it is under stress it cannot shed — heat, light, or a root zone that cannot keep up. Removing leaves does not remove the stress; it removes the plant’s capacity to deal with it, and it exposes the leaves underneath to the same lamp that folded the ones on top. Fix the cause first — distance, air movement, watering — and let the next flush of leaves come in flat. If the canopy is genuinely too dense to move air through, thin it, but do it after the tops have stopped folding, not as the treatment for them.
The half of this nobody mentions: air that is not moving
Everyone reaches for lamp height. Fewer people check whether air is actually moving across the leaves, and it changes the answer at the same lamp and the same distance.
A leaf sheds heat into a thin skin of air sitting on its surface. If that air is still, it warms up, and the leaf loses its ability to dump heat into it. Move air across the leaf and that skin thins out, and cooling resumes.
The canopy chapter of Cannabis in Controlled Environment Production (Matzneller, Gutierrez & Caplan) puts numbers on how quickly this falls off. Citing Thom (1971), it describes air moving at 2.5 m/s well above a crop dropping to roughly 0.9 m/s at the top of the canopy and about 0.25 m/s partway down inside it. The air in your room is not the air on your leaf.
The same chapter puts the useful target for thinning that layer at roughly 0.25–0.5 m/s within the canopy, citing Nelson (2011), with its authors preferring more than that where fungal pressure is a concern. In practical terms: a dense, untouched canopy with leaves stacked flat on each other is holding still air against every one of those surfaces, and an oscillating fan moving air across the tops — not blasting straight down at them — is a real lever that costs nothing to try.
We have not read Thom (1971) or Nelson (2011) ourselves. We are citing them the way the chapter does, and saying so.
Why the first instinct is to feed it or flush it
A curled leaf looks like a sick leaf, and the shelf full of bottles is right there. So the two usual reactions are to add something — often calcium and magnesium — or to strip everything out and run plain water for several days.
Both change the root zone. Neither changes the amount of energy landing on that leaf.
And there is a cost to the second one that is easy to miss. Several days on plain water in flower will genuinely start to fade a plant, and once it does you are reading a real deficiency layered on top of whatever you started with — with no way to tell which of the two you are now looking at. You have not run a test. You have added a variable.
To be straight about the evidence here: we know of no published work showing that withholding nutrients corrects a heat-driven leaf response, and we are not aware of any that has measured the harm either. What we can tell you is that it does not address the mechanism above, and that it makes the next observation harder to interpret. On the related question of whether flushing does anything at harvest time, there is a fuller answer with the actual evidence over here.
What would show that we have this wrong
If the shaded lower leaves start folding too over the following days — the ones nowhere near the lamp — then this is not a light problem and the root zone goes back on the table. Same if they never relax in the dark. And if the plant is discoloured all over rather than curled at the top, this page is the wrong page; start with the veins instead.
Where our device fits, stated narrowly
The reason this is hard to call by eye is that the number that would settle it — how hot the leaf itself actually is — is the one number a tent thermometer cannot give you. Croplock reads leaf-surface temperature from an infrared sensor in the canopy and steers on that reading rather than on a VPD worked out from the air. When it cannot get a leaf reading it trusts, it falls back to air and records that it did so, instead of inventing a leaf it cannot see.
That is a measurement, not a diagnosis. It does not tell you your leaves are curling; you can see that. It tells you whether the surface they are curling on is running hot, which is the part you otherwise have to infer. The tests above are the version you can run today with no equipment at all, and they are most of the way there.
Sources
- Nelson, J.A. & Bugbee, B. (2015). Analysis of environmental effects on leaf temperature under sunlight, high pressure sodium and light emitting diodes. PLOS ONE 10(10): e0138930.
- Matzneller, P., Gutierrez, G. & Caplan, D. “Canopy Management.” Ch. 7 in Cannabis in Controlled Environment Production (CRC Press). DOI 10.1201/9781003150442.
- Thom (1971) and Nelson (2011), on air-movement gradients into a canopy and on target air speed — both cited as they appear in the chapter above, not read in the original.
More of these at Answers. The device is at croplock.com.
From the build: Trained on close-ups, served whole plants — how the on-device classifier was built, and why it reports a family of causes rather than a single one.
Also read: How much light does it need? · Does VPD need leaf temperature? · Overwatering or underwatering? · Work out your own VPD