The Floor at Brelling
The Survey's calibration reference has read 1.83 degrees at every visit since 1974. Eleven instruments, four generations of hardware, every season. That is the problem.
Every measuring programme needs somewhere that does not change. You take the instrument there, you read the number, and if the number is the number you have got, and if it is not you have a fault to find before you go anywhere that matters. The place matters less than its stability. Ours is a wheel pit in a mill that stopped working in 1968.
The Survey has taken 341 readings at Brelling. This file is about the fact that they are all the same reading.
What a floor is supposed to do
A noise floor is not a constant. It is the residue of everything in a building that is too small to be an event: wiring, appliances, the earth's own diurnal variation, a passing vehicle, the operator's own hands. It breathes. A room measured in January and again in July will differ, usually by a few tenths of a degree, and the difference is information — it is how the site register earns its keep, and it is how you notice that somebody has installed a substation four hundred metres away.
The Survey's other three reference sites behave exactly like that. Their floors have moved, in one case by a factor of two, and every movement has an explanation in the file next to it.
| Time | Instrument | Reading | Entry |
|---|---|---|---|
| 4 May 1974 | DM-1 | 1.83° | First reading. Site selected for stability and for being dry. |
| 21 Jan 1976 | DM-1 | 1.83° | Snow on the ground, leat frozen. |
| 8 Aug 1976 | DM-1 | 1.83° | Drought. Leat dry, wheel pit dry for the first time on record. |
| 17 Sep 1983 | DM-2 | 1.83° | New instrument, new sensor, new operator. First reading taken before the DM-1 was brought out of the van. |
| 3 Mar 1991 | DM-2 | 1.83° | Second reading taken at 03:00 to catch the diurnal minimum. No difference. |
| 12 Jun 1998 | DM-3 | 1.83° | Four months after a house was built 40 m north of the mill, mains service run along the leat bank. |
| 12 Jun 1998 | DM-3 | 2.61° | The mill floor above the wheel pit, same visit. Was 2.20° in 1996. The house is visible in this figure. |
| 29 Oct 2011 | DM-4 | 1.83° | Digital instrument, first visit. Reported to three places as 1.830°. |
| 14 Jul 2022 | DM-4 | 1.83° | Drought, pit dry, ambient 34 °C. Instrument logged 40 minutes continuous; range across the window 1.829–1.831°. |
| 11 Apr 2026 | 01 | 1.83° | Most recent. Taken twice, by two operators, on two devices. |
Selected readings. The full series of 341 is in the standing register and is the most-consulted document in the archive. The mean of all 341 is 1.8300; the standard deviation, computed in 2019 by an operator who did not believe the mean, is 0.0000 to four places.
What the Survey has ruled out
The instrument. Eleven instruments have read it. The DM-1 was a moving-coil movement and a hand-wound sense coil; the current one is a solid-state part on a circuit board designed forty years later. They share no component, no manufacturer and no principle of construction beyond the physics. A shared fault is not available to us.
The operator. Readings have been taken by fifty-two people, several of whom did not know what number they were supposed to get. Since 1991 the Brelling reading has been taken blind where practical: the operator reads the instrument and writes the figure before opening the register.
The site being genuinely quiet. This is the one people reach for and it is the one the figures refuse. A very quiet site gives you a low floor that still moves. Brelling's floor is not especially low — 1.83° is unremarkable, and there are eleven sites on the register that read lower. It is not quiet. It is fixed.
The rest of the mill. The floor on the mill floor, four metres above the pit, has moved from 2.20° to 2.61° since 1996, exactly as you would expect of a building with a new house and a new mains service near it. The wheel pit, which is below it and closer to the new service, has not moved at all.
The recommendation, and what happened to it
In 2019 the reviewing officer recommended that Brelling be withdrawn as a calibration reference. The argument is sound and takes one sentence: a reference site is supposed to be a known quantity, and a site the Survey cannot explain is not a known quantity, whatever its figures look like.
The recommendation was accepted in principle in 2019 and has not been implemented, for the reason given in the minute, which is quoted here in full because no summary of it is better than it is:
Brelling is to be retained as the primary calibration reference until a replacement site demonstrates equivalent stability. The Committee notes that no site in the register demonstrates equivalent stability, and that this is the whole of the difficulty.
Cause not established.
The wheel-pit floor at Brelling has been recorded as 1.83° on every visit for fifty-two years, across eleven instruments sharing no components. The Survey has been unable to establish why it does not vary, and has been unable to establish that it should not.
The Thinplace Survey is invented and so is this file. The case files are horror fiction written in the format of a field report. The instruments they describe are real and are explained in full; the incidents are not.
Fiction