# H01W February tonal-family cross-time recurrence v1

## Decision

None of the three fixed August-discovered tonal families meets the frozen cross-time and all-node recurrence rule in this February development cohort.

The target frequencies (24.25, 31.25, and 48.0 Hz) were discovered from exposed August outcomes. The February 11--12 cohort and all processing rules were frozen before these February narrowband values were opened, but this remains an exposed-development cross-time test rather than confirmation.

## Frozen design and access

- Exactly 47 UTC-clock-matched hours: 2026-02-11 00:00 through 2026-02-12 22:00 UTC.
- Only each hour's first ten minutes, divided into five fixed two-minute windows.
- Peak tolerance: ±0.125 Hz; minimum prominence 8 dB above a 1-Hz running-median background.
- A node persists in an hour with hits in at least 4/5 windows; an hour persists only when all three H01W nodes pass that rule.
- A family recurs with at least three persistent hours spanning both UTC dates and at least two six-hour development blocks.
- All eight overlapping V8 scenes are development. The entire February 28 validator/reserve day stayed sealed.

No suitable retained product covered the fixed hours with the required frequency and node-window detail. The fallback streamed 235 exact two-minute raw-count slices and retained summaries only. No traffic, propagation, calibration, identity, or protected scene was opened.

QC passed 47/47 hours (24 on February 11 and 23 on February 12); the frozen QC gate passed.

## Family results

| Frequency (Hz) | All-node persistent hours | UTC dates | Six-hour blocks | Status |
|---:|---:|---:|---:|---|
| 24.25 | 0 | 0 | 0 | NOT_OBSERVED_UNDER_RULE |
| 31.25 | 0 | 0 | 0 | NOT_OBSERVED_UNDER_RULE |
| 48 | 0 | 0 | 0 | NOT_OBSERVED_UNDER_RULE |

No family persisted even on a single node in any hour; therefore the negative all-node result is not merely a two-node near miss. There were isolated node-window hits (24.25 Hz: 66/705; 31.25 Hz: 83/705; 48 Hz: 80/705), but the largest per-hour support was only 4/15, 6/15, and 4/15 node-windows respectively, below the frozen four-of-five rule on every individual node. These counts are descriptive support under the unchanged rule, not alternate endpoints.

These are repeated or absent narrowband morphologies under one fixed raw-count rule. They do not identify a source, detect shipping, establish attribution or causality, provide calibrated levels, validate a propagation mechanism, or support prediction. Failure is local to this cohort and rule; recurrence is not confirmation.