Two percent is a starting point: brine by weight for a hot flat

i. Why spoons stopped being good enough

For the first year I measured salt with a tablespoon, the way I had been taught in restaurant kitchens where nobody fermented anything for longer than a weekend. The trouble is that salt varies enormously by volume. The fine sea salt from the market is dense and packs into a spoon, the flaky salt a friend brought back from the Delta de l'Ebre is mostly air, and the coarse grey salt I use for olives sits somewhere between the two. A level tablespoon of each weighed eighteen, nine and fourteen grams on my kitchen scale, which is the difference between a crisp carrot and a mushy one.

A cheap scale that reads to the gram fixed this in a week. I now weigh the water into the jug, weigh the salt against it, and write both on a strip of masking tape before anything goes into a jar. It takes a minute longer than a spoon and it is the only change I have made that improved every single batch afterwards, not only the ones I was worried about.

ii. Percent of what, exactly

The confusion that cost me most was not the salt itself but what the percentage is a percentage of. Some people mean grams of salt per hundred grams of water. Others mean salt as a share of the combined weight of water and vegetables, which comes out lower for the same jar because the vegetables are counted too. Both methods work, but mixing them up between two recipes gives you a jar that is half as salty as you think.

I settled on salt against water only, because it is simpler at the counter and because I pour brine over whatever fits in the jar rather than weighing vegetables to fill a quota. For shredded cabbage and anything dry-salted I switch method and weigh salt against the vegetable, at two percent, because there is no water to measure. The masking tape always says which, with a W or a V after the figure, and the habit has already saved me twice when I went back to repeat a batch from the previous spring.

• 20 g salt to 1 litre of water is two percent brine, measured against the water
• 35 g to the litre is where I start any batch that will sit out through July
• 50 g to the litre for whole peppers and anything I want to keep firm for months
• Dry-salted cabbage: 2 percent of the trimmed cabbage weight, weighed after coring
• Olives after their water cure: 80 to 100 g to the litre, a different animal altogether

iii. What the heat does to the figure

Warmth speeds everything. A two percent jar of carrot sticks started in February, when the kitchen sits at about seventeen degrees, is pleasantly sour in ten days and stays crisp for weeks. The same jar started in the second week of July is sour in three days, soft in eight and smells of something I would not serve by the end of the second week. The bacteria I want are working fast, but so is everything else that likes a warm, lightly salted liquid, and the vegetable's own enzymes soften it whatever the bacteria do.

Raising the salt slows the whole process down and gives the good bacteria a longer head start over the moulds and yeasts. My summer figure for most batches is now three and a half percent, and for anything that will stay on the balcony rather than in the kitchen cupboard it is four. The flavour is saltier, which I deal with by rinsing a few pieces before they go on a plate, and the texture is far better. I would rather rinse a carrot than eat a soft one.

One summer I tried going the other way, lower salt and a shorter ferment moved to the fridge after three days. It works if you are home every day and have space in the fridge, and I rarely have both. The higher-salt approach suits a flat where the fridge is small and the jars have to look after themselves while I am at work.

iv. The jar card I fill in now

Every jar gets a strip of tape with five things on it: the date, the vegetable weight, the water weight, the salt weight with its W or V, and the temperature on the balcony thermometer that morning. Once the jar is open I add the day I first tasted it and one word for the texture. That single word, crisp or yielding or soft, turned out to be the most useful thing in the whole notebook, because it lines up so neatly against the temperature column.

The tape goes into the notebook when the jar is finished, stuck on the page beside the write-up, which is why some of the older pages are stiff with it. Looking back through three summers of them, the pattern is clear enough that I stopped arguing with it. Below twenty degrees two percent is fine, between twenty and twenty-six I use three, and above that I use four, check the jar every morning before work, and move anything that is ahead of itself into the fridge without waiting to see what it does next.