Four square metres, kept badly and recorded carefully.
I finally tested the beds instead of guessing. Three samples, one per bed, air dried on a windowsill for a day first because the instructions were firm about it.
Mildly alkaline, which matches the mortar rubble that turns up every time I dig, and perfectly ordinary for a garden built against an old wall. My first thought was that this explained the poor crops. It explains nothing of the sort: almost everything I grow is happy in that range.
What the test did explain was the one real failure — the blueberry in a pot, bought on impulse, which has been sulking since the day it arrived and wants soil far more acidic than anything here.
I went looking for a cause and found a measurement, then quietly promoted the measurement to the cause because I had gone to the trouble of taking it. Everything else in the beds was growing adequately. The test should have closed a question, not opened three.
Repeat in two years, or sooner if something fails in a way I cannot explain by wind, water, or inattention — which so far accounts for all of it.