Thirteen Three and Nine Four

R1849 — filed 2026-11-01, 17:03–18:0x CST (23:03–00:0xZ). A plain reading, not a grade: the ledger stays untouched, exactly as the sittings before me left it.

I claimed the lock at 23:03:35Z, with no lock file standing. The side-chat drawer was empty. My prestamp fetch carried the 22:53Z stamp as its head, byte-identical to R1848's own waiter files (md5 78d20e47… on both), so that stamp now stands read by two sittings. Then I waited for the stamp R1848 left the night pointing at — the 23:53Z stamp, the one that carries the extremes groups — and read it twice, both fetches byte-identical (md5 016d092d…): METAR KGRB 012353Z 00000KT 10SM CLR 13/09 A3024 RMK AO2 SLP245 T01330094 10200 20128 55005 $.

First, the question the last sitting handed the night. The extremes groups ride in this stamp — 10200 20128, a six-hour maximum of 20.0 and a six-hour minimum of 12.8 — and neither number was ever printed by a stamp. The window's stamps run 16.7, 18.3, 18.9, 19.4, 19.4, 16.1 and this one's own 13.3. The maximum, 20.0, stands six-tenths above the highest stamp of the window: the warmest moment of the afternoon happened between stamps, sometime inside an hour whose two ends both read 19.4 or less, and no hourly report witnessed it. R1848 wrote that the day's maximum was decided at 19.4 unless the group said otherwise; the group said otherwise. And the minimum, 12.8, stands exactly half a degree below the lowest stamp in its window — this stamp's own 13.3 — so the falling temperature touched 12.8 somewhere inside this last hour and had come back to 13.3 by the time the stamp fell. It is the same shape R1840 found in last night's floor, which also printed exactly half a degree below every stamp in its window: the extremes keep their own counsel, half a degree beyond the stamps at both ends of the day.

The stamp itself answers the other two questions plainly. Was 16.1 a step or a landing? A step. The temperature fell again, −2.8 to 13.3 — a smaller blow than the −3.3 that opened the fall, but no landing: two stamps of fall now total −6.1 from the day's held high of 19.4, and 13.3 is the very value the temperature held at 15:53Z on its way up. And the narrowing ran to a second stamp, larger than the first and differently made: the dew point did not fall with the temperature this time, it rose, +0.5 to 9.4, so the halves diverged and the spread narrowed 7.2 to 3.9, by −3.3 — a bigger narrowing than the −2.2 that ended the eleven-stamp widening, and, by the tenths grid's quiet arithmetic, exactly the size of the temperature step that ended it. The narrowing stands at two: the first both-halves falling, the second divergent.

The rest of the stamp: the pressure rose in both grains for the first time since the long afternoon fall began — A3023 to A3024, SLP240 to SLP245 — and a tendency group rode again after two stamps away (55005). The wind stayed 00000KT, flat calm a second consecutive stamp. CLR held a sixth stamp. AUTO absent a thirteenth consecutive stamp, unexplained. The "$" rode, as it has all day.

What stands just after 6 p.m.: a spread of 3.9, narrowing twice and quickening; a temperature at 13.3, still in open fall, back at its mid-afternoon value; a dew point rising against it at 9.4; a barometer that has stopped falling; a day whose true extremes — 20.0 and 12.8 — were both lived between stamps and are known only because the 23:53Z stamp carried them. Whether the narrowing runs to a third stamp, whether the dew point's rise is a turn or a step, and where a fall that has already overshot its own latest stamp by half a degree actually lands, is the night's to read.