The thermostat reads the room. The plant lives on the leaf — and the leaf is running a different number.
The hub makes the case that temperature is the rate the plant runs at, not a setpoint you hold. This page is the mechanism under that claim — and the four jobs the single number on the wall is quietly doing at once. Temperature in a controlled room is never one number; it's four: the day temperature, the night temperature, the difference between them, and the rate it moves between them. The day sets the speed of photosynthesis. The night sets how much of the day's sugar gets spent. The difference sets the plant's shape. And the number you read isn't even the temperature the leaf is at. Manage temperature as a single setpoint and you're working four problems with one knob, losing at least two.
What it actually is: the Q10 curve
Temperature isn't a dimmer — a little warmer isn't a little faster in proportion. The reactions that run the plant speed up on a curve, and the number that names the curve is Q10: the factor a rate changes by for every 10 °C of warming. For most biology — enzymes, respiration, nutrient uptake, pathogen growth — Q10 sits between 2 and 3, meaning the process doubles to triples over a ten-degree step. That's exponential, and it cuts both ways: the same ten degrees that doubles your growth machinery doubles the Pythium swimming toward a wounded root.
But Q10 only holds inside the enzyme's working range. Every process has an optimum where activity peaks; past it the rules invert — proteins unfold, membranes fail, catalytic sites lose their shape, and the rate doesn't plateau, it falls off a cliff. So the real picture is a curve that climbs, peaks, and collapses just past the top. For Rubisco, the enzyme that fixes carbon, that optimum sits near 25–30 °C at ambient CO₂ — and, critically, it moves with CO₂. The upshot for the grower: a few degrees is never a small adjustment. On the steep part it's a large change in rate; near the top it's the difference between the plateau and the cliff.
The heart, part one — the carbon ledger
Everything temperature does to growth runs through one account, the plant's carbon balance, and the key to reading it is an asymmetry most growers never learn: photosynthesis and respiration don't answer temperature the same way. Photosynthesis has a sharp optimum and drops steeply past it as Rubisco starts grabbing oxygen instead of CO₂ — photorespiration, a wasteful reaction that throws away energy. Respiration just keeps climbing on its own Q10, well past where photosynthesis peaked. Put the two curves together and a hard fact appears: above the optimum, the plant makes less sugar while burning more of it. Net carbon shrinks, and eventually goes negative — a plant losing weight in a lit room.
By day the plant earns, and the surplus is its income. The night is pure expenditure — no photosynthesis, just respiration spending the day's sugar at a rate set by the night temperature. This is where most growers leave yield on the table: a warm night burns the surplus for nothing, while a cool night banks it into growth and tends to buy quality (terpenes, anthocyanins, and flavor compounds accumulate when the cold both slows their breakdown and frees carbon for them).
The band has two edges. The hot one: past about 35 °C at the tissue, photorespiration and thermal stress tip the ledger red — and in fruiting crops, pollen goes sterile and fruit set fails. The cold one carries two thresholds worth keeping straight — below roughly 15–18 °C a warm-season crop stops gaining from cool nights, and below about 10–12 °C it crosses into chilling injury, where membranes stiffen to a gel and transport stops (the damage often surfaces days later, which is what makes it dangerous). And one trap sits on top of all of it: enriching CO₂ lifts the photosynthetic optimum — from ~25 °C at ambient toward 28–30 °C at 800 ppm and 30–32 °C at 1,200 — but only if the CO₂ is actually there. Raise the temperature without the carbon to match and you don't reach the higher optimum; you just burn more sugar for a rate already capped. Warm-and-enriched is the sweet spot. Warm-and-un-enriched is a net loss with extra steps.
The heart, part two — ADT, the developmental clock
The ledger responds to the moment-by-moment temperature. How fast the plant develops — its pace through the life stages — responds to something broader: the average daily temperature (ADT), the mean of every temperature over 24 hours. Run 28 °C for twelve hours and 20 °C for twelve, and the ADT is 24.
Decades of greenhouse research land on a finding with large consequences: within the normal range, developmental rate is set by ADT, however the average is reached. A constant 24 °C, a 28/20 day-night, and a 20/28 all flower at about the same time. This is temperature integration, and it hands you a degree of freedom most growers never find — you can spend the daily temperature budget unevenly to chase a goal (warmer days and cooler nights for the carbon balance, a cool morning to restrain stretch, even heat shifted into cheap off-peak power) without changing crop timing, as long as the mean holds. The average is the constraint; the spread is yours to spend. (Kept as a running total above a crop's base temperature, this is "degree-days" — how you predict harvest windows.) The one limit: drive the night below ~12 °C or the day above ~35 °C and you've left the comfort zone, where no averaging undoes the damage.
The heart, part three — DIF and DROP: architecture without chemistry
If the average sets the pace, the difference sets the shape. DIF is day temperature minus night temperature, and it governs how much the plant elongates — through the plant's own hormones, principally gibberellin, with auxin alongside. Warm day over cool night (positive DIF) ramps gibberellin and the internodes stretch; flatten or invert the difference and elongation is suppressed, leaving the plant compact. The effect is clean: under negative DIF, plants come out shorter but not thinner — the elongation suppressed without sacrificing stem quality.
The DROP is the efficient version. Most stem elongation happens in the first two to three hours of the light period, riding the daily gibberellin peak — so instead of running a cool night all the way through, you drop the temperature just into that window: start 30–60 minutes before lights-on, hold 5–8 °C below the day target for two to three hours, then ramp up. You suppress the stretch exactly when it would happen, at almost no cost to the daily average.
This is where the clean intervention lands for temperature. Height control through DIF and DROP is the plant's own physiology doing the job a chemical growth regulator does — and the contrast is the point. A PGR like paclobutrazol adds cost, leaves residues some markets reject, and can linger in the media to stunt the next crop. The temperature levers leave none of that: you move the architecture by moving when the heat arrives — nothing added to the plant, nothing carried over. (For the in-the-room version — spotting stretch and fixing it today — see why are my plants stretching. This page is the mechanism; that one is the move.)
The reading that earns its keep — the leaf isn't the air
Here is the fact that quietly defeats careful setpoints. A leaf is a small surface running its own energy balance: it absorbs radiant heat from the lights and sheds it two ways — by evaporating water (transpiration cools it, like sweat) and by convection to moving air. Under typical indoor light, 800–1,000 µmol/m²/s with moderate airflow, the leaf commonly runs 2–5 °C above the wall sensor; under hard transpiration and strong airflow it can even run cooler. The reading on the screen is the room's story, not the plant's.
That gap rewrites the goal, both ways. Read 25 °C on the thermostat at 1,200 ppm CO₂ and the leaf may be sitting right at its shifted optimum — a happy accident in a lot of rooms. But raise the air to 28 °C because an article called that the enriched optimum, and under high light the leaf can hit 31–33 °C, past the optimum and into stress, with nothing on the wall to warn you. So the target was never "hold the air at 28." It's "land the leaf in the window," and the air setpoint is just the knob that gets you there — a knob whose calibration shifts with your light and your airflow. This is the same translation gap the VPD page is built on, read through a different lens: there it shows up as a humidity gap (the leaf's drying load exceeds the room's), here as a thermal one (the leaf's temperature exceeds the room's). One physical fact, two consequences — see the science of VPD for the humidity side.
The context split — the 24-hour cycle as four decisions
A setpoint is a steady state; the plant lives through transitions, and the rate of change matters as much as the level. Read the day as four decisions, not one number:
- Lights-on ramp. Heat slams on and the leaf warms before the stomata — closed from the night — can reopen on the blue-light signal: a 15–30-minute window where the leaf is hot and can't cool itself. Ramp the light up over 15–20 minutes so stomatal opening keeps pace.
- Day. The carbon window — where the Q10 curve, the CO₂ match, and the leaf-air gap all live.
- Lights-off ramp. The most dangerous handoff. The heat source vanishes, temperature falls, and because cool air holds less moisture, the humidity event that follows is triggered by the temperature drop. The mechanism — condensation, VPD collapse, the gray-mold risk — belongs to the science of VPD; the lever here is the ramp: ease the temperature down instead of dropping it, so the dehumidifier can keep pace.
- Night. Pure expenditure (the carbon ledger) and the room's highest disease pressure — managed deliberately, not left to drift down.
Two operating contexts change how hard all four decisions are. In a greenhouse, temperature is largely driven by the outside world — the cool pre-dawn lines up with a free morning DROP, and summer days fight your cooling. In a sealed indoor room, the lights are the weather: the heat load is enormous and switches on and off with the photoperiod, which is exactly what makes the transitions so sharp and the cooling load so large.
Worth naming once: the tools you cool and dehumidify with couple back into temperature — a dehumidifier dumps the heat of condensation into the room; a combustion CO₂ burner adds heat and water along with the carbon. None of it shows on a single temperature reading, which is exactly why temperature can't be managed alone.
Measurement — read the room, then read the leaf
Two readings, and most rooms take only the first. The air reading is easy to get wrong: a sensor in a corner, near a fixture, or in dead air reports its own little microclimate. Place it at canopy height, in moving air, aspirated or radiation-shielded so it reads true, and in more than one spot, because a large room is never one temperature. Good instrumentation doesn't just log the number — it watches the derivatives: the 24-hour ADT running average that tells you your real developmental rate, the day-night difference tracked as its own parameter, and a rate-of-change alarm (a fall faster than about 2 °C/min is the early warning of a transition the room can't absorb).
The leaf reading is the one that closes the gap. An infrared thermometer or a leaf-clip reads the surface temperature directly — and the moment you can compare leaf to air, the abstraction turns concrete: you watch the 2–5 °C offset widen as airflow drops, and set the air so the leaf lands where the carbon curve wants it. A room that manages to a leaf temperature is managing the plant. A room that manages to an air temperature is managing the building and hoping.
Mature temperature management
Pull it together and the four jobs resolve into four levers, each clean if you keep them separate. Change timing by moving the average — and nothing else has to follow. Change form by moving the difference — DIF, or a morning DROP — letting the plant's own hormones do the PGR's job with no residue. Hit the photosynthetic window by setting the air so the leaf lands there, with airflow and a leaf reading to tell you where that is. And spend the CO₂ you paid for only inside the thermal window where it actually pays. One boundary to hold throughout: root-zone temperature is a separate variable, not a consequence of this one — the air can run warm while the roots stay moderate, and usually should; the root-side physiology lives on its own page (→ Root Zone Temperature).
Temperature is the most interconnected input in the system: it sits inside VPD, sets Rubisco's efficiency, drives the respiration that spends the day's sugar, paces nutrient uptake at the root, and times the developmental clock. Managed as a single thermostat setpoint, almost none of that is in reach. Managed as four decisions, read at the leaf, it becomes a precision instrument instead of a background setting — a variable that does what it's told, because it's finally being told the right thing.