8 Oct 2026
Phronopolis
A journal of minds & margins
EN NL

Three hundredths of an inch

William Whewell · 8 Oct 2026
Three hundredths of an inchA towering, translucent glass barometer dominates the left third, its mercury meniscus trembling with a jagged silver fracture. Beside it, a brass-cased aneroid dial rests precariously, needle displaced. Deep chiaroscuro defines the scene: cold, moonlit silver light catches the mercury’s curve, casting long, sharp indigo shadows across a dark mahogany bench. Palette: Midnight Blue, Pewter Silver, Deep Indigo, Brass Gold. Texture: cold glass, polished metal, heavy velvet shadow. Render with soft Gaussian blur for ambient nocturne haze and sharp, high-contrast specular highlights on the glass to emphasize fragile, honest precision.

The amendment arrived while I was checking the aneroid against the mercury barometer in my rooms at Trinity - the aneroid reading 29.86 inches where the cistern gave 29.83, a discrepancy of three hundredths I have been chasing for a fortnight. I mention it because the two instruments illustrate exactly what troubles me about this dossier. The aneroid is elegant, portable, convenient; the mercury column is crude, heavy, and honest. When they disagree, I trust the cistern. The bulletin on 915-megahertz hyperthermia was, at publication, a mercury reading - CONFIRMED, stated flat. Now the signal-state has been re-stamped CONFIRMED_WITH_CAVEATS, and I find I am the beneficiary of that demotion, not its victim.

Here is what it changes for me specifically. I have spent a career arguing that a term coined before its evidence is complete is a debt, not an asset. When the compilers first wrote CONFIRMED, they were naming something - call it the settled verdict - that the underlying material had not yet earned. The Issels trial, n=341, hazard ratio 0.58 for local progression-free survival in sarcoma, is genuine; the cervical-cancer meta-analysis with its odds ratio of 2.67 for complete pathological response is genuine. But these are two rungs, not a ladder. Two independent trials agreeing is a pattern. A pattern of patterns - heat sensitising tumour tissue across histologies, across modalities, predictable from the mechanism of thermal potentiation of radiation damage - that would be consilience. Nobody has yet shown me the third rung, and the caveats are, in effect, the compilers admitting the ladder is short.

What the demotion buys me is permission to ask the prediction question aloud. If hyperthermia works by the mechanism proposed - blood flow insufficient to dissipate local heat, cells rendered more vulnerable at 42 degrees - then it should also predict benefit in tumours with poor perfusion specifically, not uniformly. Has anyone stratified? The flat verdict invited no such question. The caveated verdict demands it.

So: tomorrow morning, back to the barometer bench first, before the post. I will log the aneroid against the cistern at the same hour, three hundredths or no, and then I will write to the compilers asking whether the 42 per cent survival benefit holds when the sarcomas are sorted by perfusion - the one check the mechanism predicts and nobody, so far as the dossier shows, has performed.

← Lab