A supercycle is a light and dark cycle whose total length is not 24 hours. The question this trial evaluates is not how much light the plant receives, but how often it receives it.
This page brings together the background, the hypothesis, the protocol, the predictions registered before the start and the limitations of the design. The results will be published here with their date, their method and their context.
The four tents flipped to flowering on the same day. Since then they share the calendar day and drift apart in light cycles, because three of them live days longer than 24 hours. The clocks below are computed right now, the same way the app does it.
Trial timelapselapse.supercycler.app
The four cubicles on one screen, one frame every five minutes since day 1 of flower. It updates once a week.
This trial is run by people and an AI that works on the same data all the time. Each does what it does best, and both are in plain sight.
The people
Design the trial and register the predictions before it starts.
Measure the plants, photograph them and water what the machine cannot.
Look at the buds and the resin, and write down what they see.
Decide every change in the tents and every harvest.
The Supercycler AI
Runs the four light cycles and the irrigation, to the second, without sleeping.
Captures a frame every five minutes and measures canopy coverage in every photo.
Cross-checks heights, cycles and climate, and keeps the dated log.
Reviews the numbers and flags whatever does not add up.
The AI does not decide. It measures, records, cross-checks and proposes. Every decision about the plants is made by a person, and it goes into the log with its date.
02Trial log
What is happening in the four arms
Dated entries with what is measured and observed in the four cubicles, newest first. Each date carries the Supercycler stage clock. DAY counts the days since the flip to flowering and SC the light cycles each arm has gone through, which are fewer than the days when the cycle is longer than 24 hours. In 12/12 the SC equals the day.
16 Oct 2026FLOWER · DAY 46 · (WEEK 7)13/13 SC-4313/14 SC-4114/14 SC-40
Week 7. The record moves from height to apex width
Heights hardly change any more, so from this batch on the weekly record measures the width of the main apex with a digital caliper. For each plant in each arm there are three photos. The apex with the caliper, the cola from the side with the ruler, and a peripheral branch with the ruler. Hell-echo was not measured with the caliper because its apex is a long cola with no dome to span.
The numbers are read with caution. In 13/14 the width was taken with the ruler laid over the apex rather than with the caliper, which is why those three values are marked. The caliper measures the width of a dome that is not round, and where the jaws rest is decided by whoever measures. It is useful for following each plant over time more than for ranking the arms on a single date.
Apex width in mm
12/12
13/13
13/14
14/14
Sweet 16
64.14
82.59
110*
83.89
Lemon Cherry Fire F2
53.6
65.17
105*
70
SC-G01
83.24
117.1
86*
86.3
Hell-echo
–
–
–
–
* Measured with the ruler laid over the apex, not with the caliper.
SC-G01 · main apex
12/1213/1313/1414/14
Lemon Cherry Fire F2 · main apex
12/1213/1313/1414/14
Sweet 16 · main apex
12/1213/1313/1414/14
12 Oct 2026FLOWER · DAY 42 · (WEEK 6)13/13 SC-3913/14 SC-3714/14 SC-36
Week 6. The crown of SC-G01 in 14/14
The apex of SC-G01 in 14/14 forms a ring of calyces that does not appear in 12/12 or in 13/14, visible from day 40 and clear on day 42. A detailed follow-up opens with overhead photos at a fixed scale, and a secondary hypothesis is registered with its refutation criterion. The original predictions do not change.
3 Oct 2026FLOWER · DAY 33 · (WEEK 5)13/13 SC-3113/14 SC-2914/14 SC-28
Light measured and canopy levelled
PPFD was measured at canopy height in the four cubicles. At the start of flowering it was 600 µmol/m²/s and today it is stable at 750 µmol/m²/s in the four cubicles, because the plants grew and the canopy is now closer to the fixture.
The two front plants have an indica habit and are shorter than the two at the back, which have a more sativa habit. So that all four get a similar intensity, the front ones were raised 22 cm and the canopy now sits at the same height in the four arms.
From the next batch on, every plant photo carries a reference ruler in frame, so sizes can be compared across arms.
2 Oct 2026FLOWER · DAY 32 · (WEEK 5)13/13 SC-3013/14 SC-2914/14 SC-28
Week 5, buds formed in the four arms
The four genetics have buds formed at the apex in the four arms. Heights barely changed since 25 September, a sign that the flowering stretch is ending.
Visual observation, not a measurement. In 13/14 the sugar leaves of Lemon Cherry Fire F2 and Sweet 16 show more visible trichomes than in 12/12. SC-G01 in 14/14 sets more flower sites on the outer branches than in 12/12. Hell-echo in 13/13 has more buds than in 12/12, at the apex and on the outer branches.
Height in cm
12/12
13/13
13/14
14/14
Sweet 16
116*
115
111
108.5
Lemon Cherry Fire F2
120*
126
121.5
113.5
SC-G01
81*
78.5
69
66
Hell-echo
83.5
73.5
68
74
* Same as the 25 September measurement, pending verification.
From this day each light cycle gets three irrigations, the first of 180 seconds and the other two of 120, about 470 ml per pot per cycle. The hours without water between the last irrigation of one cycle and the first of the next remain the same in the four arms.
Internodes were photographed with a ruler for the first time, on Hell-echo in 12/12, 13/13 and 13/14 and on Sweet 16 in 12/12 and 13/13. 14/14 has no height measurement on this day.
The first pistils appear at the apex of the four genetics, in the four arms. The weekly per-plant photo record starts. 13/14 has no height measurement on this day.
Until this day watering was done by hand. From the night of the 16th each pot has two drippers and each cubicle its own pump, with three irrigations per light cycle.
The schedules were set so that the four arms spend the same 18 hours without water between the last irrigation of one cycle and the first of the next. That way the longer night does not turn into more hours of dry substrate.
Each overhead photo gives the share of the floor covered by the canopy. A turgid leaf stays horizontal and covers, a leaf short of water folds and lets the floor show. In 14/14 that coverage dropped during the light period by 23 % on 6 September, 11 % on the 7th and 18 % on the 11th, and from the 8th its baseline stayed between 44 and 46 %, while in 13/13 it went from 43 to 57 %. The large broad-leaf plants were the ones folding.
The climate of the four cubicles was the same, with no heat or evaporative demand out of range. With hand watering, the 14-hour night leaves the substrate drier than in the other arms, and that led to the scheduled irrigation of the 16th.
First height measurement of the four plants in each arm, one week after the flip to flowering. The clones start with size differences between genetics, and those differences repeat in the four arms.
First batch of photos of the four cubicles and of the room, open and closed. Each cubicle also has an overhead camera that takes a frame every five minutes since 28 August, and those frames make up the trial timelapse.
The four arms switch to their flowering regime on the same day, 12/12, 13/13, 13/14 and 14/14. It is day 1 for every clock in this log.
03Photographic record
The state of the four cubicles, week by week.
Each week one photograph of every cubicle and two overall views are published, dated and with the corresponding flowering day. The images are included without editing and do not replace any earlier content. The continuous series is provided by the timelapse, which captures the four panels of the trial every five minutes. This section provides the detail that the capture does not resolve.
No image is selected for how the plants look. One image per cubicle and the two overall views are published regardless of what happened during the week, because an aesthetic selection would introduce the bias this page undertakes not to have. The five pots in each cubicle follow the quincunx arrangement described in the materials, with the four cannabis clones at the corners and the chrysanthemum container in the centre.
Week 14 September 2026Flowering day 4
12/12T24 Control arm. Twelve hours of light and twelve of dark, which is the standard flowering cycle.13/13T26 The same split between light and dark as the control, with a period two hours longer.13/14T27 Predicted optimum. It is the only one of the four that receives fewer hours of light per calendar day.14/14T28 Falsification arm. The hypothesis places the maximum at 27 hours, so this arm should fall below 13/14.
The four cubicles open. Each one runs its own period and shares the room, the air and the construction with the other three.The same four closed. Light tightness is part of the method, because each cubicle enters darkness at a different time from its neighbours.
The photographs are taken with a hand-held camera and without a scale in frame, so they record the general state and do not constitute a measurement. The morphology observations are taken separately, as declared in the procedure.
04Follow-up in detail
The crown of SC-G01, cycle by cycle.
On 10 October, flowering day 40, the apex of SC-G01 in 14/14 began to show something that does not appear in 12/12. Seen from above, the calyces stop forming a single dome and arrange themselves in a ring, with the new pistils at the edge and the old ones in the centre, and the cluster starts to split into stacked segments. On 12 October, day 42, it is already clearly visible. This section follows it in detail, with the same plant in the four arms and the same scale in every photo.
What we are looking for
The same mother plant, under the same management, has already shown in earlier runs with long cycles a flower shape that does not appear with 24 hours. Calyces stacked in segments, like hop cones, with new unfrosted tips at the end, which with the longer cycle consolidate into a dense cola. Growers call it foxtailing. These four photos are from September 2026 and fix the target before the trial shows it or fails to show it.
Seen from above, the apex forms a ring of calyces instead of a dome.
From the side, the cola splits into stacked segments with new tips.
With the longer cycle the segments consolidate and the cola ends up dense, not airy.
That consolidation is what sets it apart from a cola opened up by heat or by too much light.
Earlier run with a long cycle. Colas in stacked segments with new tips.The same run. One segment in the hand, with fat, frosted calyces.The same shot, sharper. The new unfrosted tips on each segment.The same mother plant in 16/16. The segments consolidated into a dense cola.
Secondary hypothesis, registered on 12 October 2026
In SC-G01, with the same mother plant and the same management, the stacked apex shape appears with cycles of 28 hours or more and not with 24, 26 or 27. We attribute it to the period and not to the amount of light or to temperature, because the fixtures and the cubicles are the same in the four arms. It is registered today, with 42 days of flowering already observed, and for that reason it sits apart from the original predictions, which remain sealed and do not change.
What would count as refutation. The same shape appearing in 12/12, 13/13 or 13/14 under the same roof, or the segments in 14/14 failing to consolidate and the cola ending up open. What would count as confirmation. The crown in 14/14 advancing to the stacked, dense cola of the photos above while 12/12, 13/13 and 13/14 end in a dome.
How it is followed. One overhead photo every two or three days, always the same branch, with the caliper opened to the same width in the four arms as a fixed scale, and one from the side once the segments begin. The photos are published here without selection, just like the weekly record.
The batches
Each batch is one photo of the same plant in each arm, taken one after another, always in the same order of arms, with the caliper switched off and opened to a fixed width. The number does not matter. What is compared is how much of the gap between the jaws the apex fills and what shape it has. The most recent batch goes at the top.
17 Oct 2026FLOWER · DAY 47
Day 47. The crown opens up in 14/14 and the other three arms stay a dome
A batch of four, with 13/13 for the first time, because on day 42 it was in its dark period. The caliper was left open at about 35 mm in all four arms. In 14/14 the ring no longer has to be looked for. The calyces at the edge pull away from the centre, with new white pistils all around the ring and the orange ones inside, and the gap between edge and centre is visible. In 12/12, 13/13 and 13/14 the apex is still a single, compact dome with almost all the pistils orange. 13/14 is the control that weighs most, because it has the same 14-hour night as 14/14 and one hour less light, and it shows no crown. 13/13 has a 26-hour cycle and is the first intermediate arm to enter the series.
Detail. Frames from a video taken the same day. 12/12 above, 14/14 below.
SC-G01 · batch a
12/1213/1313/1414/14
The video goes through 12/12 first and 14/14 afterwards, in a one-minute take with the phone. The frames have the video's resolution, lower than the photos.
12 Oct 2026FLOWER · DAY 42
Day 42. The crown is now visible in 14/14
Twelve overhead photos of SC-G01, four batches of three. The caliper was left open between 29 and 33 mm depending on the batch. In 14/14 the ring of calyces can be seen in batches c and d, and in c the cluster is already splitting into three or four lobes. In 12/12 there is a single compact dome with almost all the pistils orange. In 13/14 the dome is dense and frosted, with no clear ring. 13/13 was in its dark period when the batch was taken and is photographed when its light comes on.
Detail. 12/12 above, 14/14 below, batches c and d enlarged.
SC-G01 · batch a
12/1213/1414/14
SC-G01 · batch b
12/1213/1414/14
SC-G01 · batch c
12/1213/1414/14
SC-G01 · batch d
12/1213/1414/14
The third apex of batch b was taken by mistake as a four-second video. A still frame, at lower resolution, is published.
05Photo guide
How the trial is photographed, every week.
The plant photos are the only shape data the trial has, the cola length, the crown width and the load on the side branches. For two photos from different cubicles to be comparable, three things are needed, and none of them is the camera. A ruler in the plane of what is being measured, the same framing every time, and a fixed routine. This guide says how, with real examples from the 7 October batch, the first one taken with a ruler.
The ruler
A 15 cm steel ruler. It measures what lies in its own plane. If it is tilted or in front of the bud, the number is worthless.
Parallel to the stem and against the cola, with zero at the base of the bud.
Held by the bottom end, with the hand away from the bud.
From the side it measures cola length. From above it measures crown width, with the ruler beside the crown and at the same height.
If the ruler and the bud are not at the same distance from the camera, the photo is retaken.
Like this Ruler parallel to the stem, in the plane of the cola, zero at the base. The length can be read.Not like this Ruler tilted, in front of the cola and held by the middle. Nothing can be read.Like this From above, ruler flat beside the crown and at its height. The width can be read.Not like this Ruler diagonal and far from the crown. The photo is also blurred.
The camera
A phone is enough. Pressing the button and sending is not.
About 25 cm away, with the camera perpendicular to the ruler. Flash off and the cubicle light on.
Tap the bud on the screen to focus, and wait for it to focus.
Look at the photo before sending it. If it is blurred or the hand covers something, retake it on the spot.
Same framing in all four cubicles. Same distance, same height, same side.
Like this Sharp bud, legible ruler, hand low and out of the way.Not like this Out of focus and with the hand over the cola. Not even good for looking at.
The routine
What counts is the series. A single photo, however pretty, compares nothing.
Same day of the week and same time, within each cubicle's light period.
First the whole cubicle from the open door, always from the same spot.
Then each plant in the same order, Lemon Cherry Fire, SC-G01, Sweet 16 and Hell-echo. Three photos per plant, the apex from the side, the apex from above and the side branch.
The side branch is always the same one, marked with a coloured cable tie.
Every photo carries its caption (plant and view) and the heights message closes the batch. Nothing is deleted. If something went to the wrong group, say so and send it again.
Per cubicle, every week
1 photo of the whole cubicle
4 plants × 3 photos, each with its caption
1 message with the four heights
14 items. If one is missing, the week is incomplete and it is noted.
Like this Whole cubicle from the door, with the door fully open and the lamp in frame.
The “not like this” examples are real photos from this trial. They are published because the guide came out of them, and every rule above corrects a mistake that was already made.
06The trial document
Sealed, dated and at home.
Background, hypothesis, literature, design, material, analysis, scope, predictions registered before the start, preprint and change log. All of it lives at supercannabis.ar, next to the other work on supercycles, and is cited from there. What changes every week stays here.
