In a flame, a sheet less than a millimetre thick turns fuel into products. Almost all of the flame's entropy is produced there, and the record of it leaves: light escapes, products flow away. A laser crosses each room, and its absorption line is compared with a cold reference cell. Under the rate reading, the line moves with entropy produced; under the stock reading, it moves with the record that is present.
| Outcome | A | K1 | K2 | Meaning |
|---|---|---|---|---|
| Rate reading | shift | = A | ≈ 0 | Rate reading survives; stock reading fails. |
| Stock reading | shift | 0 | ≥ A | Stock reading survives; rate reading fails. |
| Line moves up in frequency | blue shift | Time-Gel's sign rule (clocks slow) fails. | ||
| No shift anywhere | 0 | 0 | 0 | Only small ρcrit values fail: the bound rises. The theory as a whole is not refuted while ρcrit has no upper limit. |
The line is drawn to scale when the shift is large enough to see; otherwise it is magnified by the factor shown. The flow law is the site's own, dτ/dt = √(1 − ρI/ρcrit), the one the live model runs. Published flame measurements we found were taken in burnt gas or calibrated against lines in the same flame, so they set no bound on the reaction zone yet. Protocol: Information Leaving Its Owner v1.0, companion to the ice clock.
Model output, not experimental evidence. No flame line has been measured for this test yet.