The first genuinely sweet cider I ever drank was not backsweetened, pasteurized, or filtered. It was French, it was about four percent alcohol, it had a faint honey edge under the tannin, and it had stopped fermenting on its own with sugar still in the bottle. I spent a season trying to reproduce it with sorbate and sugar additions before finding out the thing has a name and a long tradition behind it. It is called keeving, and it works by starving the yeast rather than killing it.
Sweetness that survives without sorbate
Every other route to a sweet cider intervenes at the end. You ferment dry and add sugar back with a stabilizer, you pasteurize the bottles, or you filter the yeast out — all of which I have done while trying to avoid bottle bombs. Keeving intervenes at the start instead. You strip most of the nitrogen out of the juice before fermentation begins, so the yeast never gets the raw material to finish the job. Fermentation crawls for months and then simply stops with residual sugar intact, and there is nothing left to stabilize.
The result is a different drink from a dry cider with sugar in it, which is the part that surprised me. A slow, nutrient-poor fermentation preserves more of the apple, so the sweetness reads as fruit rather than as an addition.
How a pectin gel takes the nitrogen with it
The mechanism is worth understanding before you try it. Andrew Lea, a food biochemist who spent thirteen years in the cider section at Long Ashton Research Station and took his PhD on apple tannins, lays it out in his write-up of what keeving actually is, and the short version runs like this.
Apple juice contains pectin. An enzyme called pectin methyl esterase, PME, de-esterifies that pectin into pectic acid. Calcium then cross-links the pectic acid into a gel. Here is the clever part: pectin carries a negative charge, while the nutrients the yeast needs most — thiamine and amino nitrogen such as asparagine — carry a positive one. As the gel forms, it pulls those nutrients in with it. Lea reports that both the pectin and the amino nitrogen come down by at least fifty percent.
Meanwhile a very slow fermentation has begun inside the gel, and the gas bubbles it throws lift the whole mass to the surface as a firm brown raft — the chapeau brun, or brown cap. Heavier solids sink. What is left between cap and lees is bright, nutrient-stripped juice, and that middle layer is what you are after.
The juice a keeve needs, and the juice that will refuse
Keeving is fussier about its raw material than any other cider technique I use. Four conditions matter.
- Enough sugar. Lea puts the floor at twelve percent sugar, around 1.055 specific gravity. Thin juice keeved and stopped early gives you something closer to a soft drink than a cider.
- Naturally low nutrient. The whole method depends on the juice starting nutrient-poor, so cider fruit from unfertilized trees keeves and heavily fertilized dessert fruit often will not. Store-bought juice is the worst candidate of all, since it has usually been enzyme-treated and has little pectin left to gel.
- The right fruit type. Bittersweets are the classic keeving apples. Lea notes that modern enzymes have brought bittersharps into range down to pH 3.6 and below, but he is equally clear that with some varieties no cap forms at all. This is not a failure of technique; it is the fruit.
- Cold, and staying cold. Around 5°C, holding there for a week or more. Keeving is an autumn-into-winter operation, which is why an unheated garage works and a kitchen cupboard does not.
If you press your own fruit, a maceration of up to twenty-four hours before pressing helps, because the pectin and the PME both need to come out of the fruit tissue and into the juice.
PME and calcium, added in that order
Lea's doses are specific, and so is the sequence. Calcium chloride goes in at 400 parts per million, which he gives as four grams per ten litres, and the PME preparation at roughly twenty millilitres per hundred litres of juice.
The sequence is the part not to improvise: the two are never added together, only as separate sequential operations. Dump calcium into juice that already has active PME in it and you will get gel where you are stirring rather than an even set through the batch.
One more warning worth repeating, because it is the mistake that a shelf of homebrew supplies invites. Keeving PME must be free of any polygalacturonase, the depolymerizing activity that ordinary pectic enzyme is sold for. The pectic enzyme you use to clear a hazy cider does the exact opposite of what a keeve needs — it chops pectin up so it cannot gel. A keeving enzyme and a clearing enzyme are not interchangeable, whatever the shelf next to them looks like.
Racking the middle layer, and the volume it costs
Timing the rack is the skill in keeving, and it is a judgment call rather than a measurement. You want to siphon just before the cap begins to break up. Leave it too long and the raft disintegrates, the nutrients it captured redissolve, and you have an ordinary cider fermentation with an expensive backstory. Cap formation takes about a week at cellar temperature, so the vessel needs looking at daily once it starts to rise.
Take the clear juice from between the cap and the lees, and expect to lose around twenty percent of your volume to the two layers you leave behind. A keeve is not the batch to run when you have pressed exactly five gallons and want five gallons of cider. It is also an argument for a taller, narrower vessel, where the same volume gives a deeper clear band in the middle.
From there the fermentation is a waiting game measured in months rather than weeks. Lea's guidance is to bottle at 1.015 or 1.010 and expect a further drop of about five gravity points as the cider conditions in the bottle. That is a real bottle-conditioning fermentation producing real pressure, so it belongs in proper pressure-rated glass with the same respect I would give any bottle-conditioned cider.
Where a home keeve goes sideways
Three failure modes cover nearly everything I have seen or done wrong.
- A white cap instead of a brown one. That is vigorous yeast growth outrunning the gel, and it means the keeve has failed rather than merely stalled. Too warm, too much nutrient, or too big a pitch of aggressive yeast.
- No cap at all. Usually the fruit, sometimes juice that was already enzyme-treated. Nothing you add later rescues it — ferment it out as a dry cider and choose different apples next season.
- A fermentation that sticks well above your target. Keeving deliberately walks up to the edge of a stuck fermentation, so it can overshoot. Lea's remedies are small and specific: thiamine at 0.2 ppm, or diammonium phosphate. Add either one and you are trading sweetness for completion, which is a decision, not a repair. My usual stuck-fermentation checklist assumes you want the sugar gone; here you mostly do not.
What I would tell someone starting their first keeve
Run it small and run it in parallel. Press your fruit, split the juice, and keeve five gallons while fermenting the rest dry from the same pressing. That gives you a control, and a failed keeve then costs a fifth of the day rather than all of it.
Then keep three numbers in a notebook: the starting gravity, the date the cap rose, and the date you racked. Those three tell you more about whether to repeat the experiment than any tasting note will, because keeving is fundamentally a timing craft and the timing is what varies with your fruit and your cellar. Get a cap, rack the middle, and be patient into spring — and the cider that comes out will be sweet because the yeast quit, which tastes noticeably different from sweet because you intervened.