There is a moment, when you start a ferment, that you have already made every important decision. Before the cabbage has begun to wilt, before the salt has begun to draw the brine, before any bacteria have noticed anything, the outcome of the next ten days has already been set. By the salt percentage you chose. By the temperature of the room you put the jar in. By the size and shape of the jar itself. By the cleanliness of your hands and the cleanliness of the vegetable. By the time of year. By the question — which you may not have asked yourself out loud — of whether the jar will be checked tomorrow, or whether the cook will travel and the jar will sit unattended.

These decisions have been made. The next ten days are mechanism unfolding from the decisions.

This is the central thing the Notebook will keep saying. Fermentation is not a flavor pursuit. It is environment design. The cook does not choose which bacteria win the cabbage by tasting and adjusting along the way. By the time taste is meaningful, the bacterial population has already been decided. The cook chooses by setting the variables before the ferment begins. Salt percentage. Temperature window. Time horizon. Oxygen access or its absence. Vessel. Substrate. After that, the jar is a small room with chemistry rules, and the cook is no longer in control of who lives there — only of who was invited.

This is uncomfortable for the home cook trained on recipes. A recipe says: cook to taste. Adjust seasoning. Bring back to a simmer. Reduce until thickened. These are the verbs of fresh cooking. They assume the cook can intervene at every step. Fermentation does not assume that. A nukazuke bed, once started at 7% salt and tended daily, does not respond to "more salt" or "adjust seasoning" in the way a soup does. The mass is too large, the time horizon too long, the ecology too settled. By month two, the population that lives in the bed is the population that will live in the bed. Adjusting the salt now is like trying to renovate a house while a family lives in it. You can do it. You will get a different house. But the family will not be the same family afterward, and that may not have been what you wanted.

The same is true at the smaller scale of a single jar of sauerkraut. By the seventh day, the lactic acid bacteria have produced enough acid that the pH is well below the threshold most other organisms can survive. The ferment has become its own ecosystem. The cook who, on day seven, decides the cabbage "needs more salt" cannot really add more salt; the cabbage is already a pickle, and the salt would be a topping on a finished product, not a variable in a decision. The window for environment design closed somewhere between day one and day three.

This is the first thing to internalize about fermentation, and almost no fermentation book says it directly. The cook is not a chef during a ferment. The cook is a gardener. And gardening has a different relationship to time than cooking does.

What this Notebook is for

This Notebook is for the cook who has done one of these:

• Started a jar of cabbage that turned into something good — and wants to repeat it, knowing why, on purpose, without luck.
• Started a jar that turned into something bad — and wants to know which of the variables they got wrong, so the next jar does not.
• Has read about fermentation in passing — in the Atlas chapter, in a magazine article, in a cookbook footnote — and wants the actual numbers. The salt percentage. The temperature range. The day-seven test. The discard rule.

The Notebook is not a recipe collection. The recipes — for sauerkraut, kimchi, nukazuke, shio-koji, vinegar pickles, miso starter — exist for free at terumimorita.com. They will continue to exist for free. What the Notebook does is sit alongside those recipes and explain what is happening in the jar, what the cook should be looking for at each stage, and which of three variables (salt, temperature, time) is responsible when something has gone unexpected.

In that sense, the Notebook is more like a manual than a cookbook. It is the operating manual for the small ecosystems the cook is choosing to start.

It is also a safety document. Of all the cooking systems a home cook can pursue, fermentation has the longest list of things that can go badly wrong if the cook ignores certain rules. Most of those rules can be stated in one sentence. The Notebook will state them in one sentence each, and then explain why, because rules that come without their reasons get forgotten, and forgotten safety rules in fermentation produce real harm.

What this Notebook is not for

This Notebook makes no claim about the health benefits of fermented food. Fermented food may or may not be good for any particular person, and the variation in salt loads, bacterial profiles, pH, and alcohol content across world fermentation traditions is large enough that single claims do not generalize. The Notebook makes no claim about gut health, immunity, longevity, or any other physiological outcome of eating any of the foods it describes. If those claims interest you, they are easy to find elsewhere. They are not what this book is.

The Notebook is also not a defence of wild fermentation. Wild fermentation — fermenting without controlled salt percentages, without temperature management, without a tested ratio — sometimes works. When it works, the cook attributes the success to something in the room, or to a general feeling that fermentation is forgiving. When it does not work, the result is a jar that smelled clearly of putrefaction within forty-eight hours, and the cook learned only that wild fermentation is a coin flip. The book the Notebook is in conversation with takes the coin flip as a teaching opportunity, not as a method. Every ratio in this Notebook is a number, not a feeling. Every temperature is a window, not an idea.

This is not because there is anything wrong with the looser tradition. It is because the cook who is reading this Notebook is most likely trying to repeat something. Wild fermentation is poor at repeating. The Notebook is for repeatable.

How to use this Notebook alongside the public recipes

The Notebook does not replace the recipes at terumimorita.com. It sits next to them. Each chapter assumes the reader has access to the corresponding recipe page and can read the ingredient list and the instructions there. What this Notebook adds, on top of that:

• The number the recipe shows you as a quantity is here shown as a ratio — usually a percentage of total weight — which is what actually matters for the ferment.
• The temperature the recipe gives as "room temperature" is here given as a window in degrees Celsius, with the upper bound at which the wrong organisms become competitive.
• The time the recipe gives as "ferment for one to two weeks" is here broken into day-by-day observations — what should be happening on day one versus day three versus day seven, and how to tell when the ferment is finished.
• The failure mode the recipe does not discuss — because recipes are usually written under the assumption that nothing will go wrong — is here listed explicitly. What it looks like. What it smells like. Whether it can be saved. The single most important question this Notebook tries to answer for the home cook is: when is it time to throw it away? The wrong answer has a higher cost than the wrong answer in any other home-cooking situation. The Notebook gives that answer directly.

If you have never made a sauerkraut and want to make one, follow the free recipe. Then come back to the Notebook before you start your second one. The second sauerkraut, made with the Notebook open, will not necessarily be more delicious than the first. It will be the same sauerkraut. But you will know which decisions you made, and you will be able to make different ones next time on purpose. That is the value the Notebook is trying to add. Not a better recipe. A clearer relationship to the jar.

The three variables that govern everything (Chapter 1 preview)

Almost everything that happens in a fermentation jar can be predicted by three variables. The remainder of this Notebook is, in a real sense, just explaining how those three variables interact with each other and with the substrate.

The three are:

• Salt percentage — the threshold variable. Determines which bacteria can compete in the brine. Set this wrong by one percentage point and the ferment may fail outright; set it right and the rest of the variables become flexible.
• Temperature — the speed variable. Determines how fast everything happens. A ferment at 22 °C and a ferment at 28 °C are running the same chemistry at different rates, with different windows for the cook to read and intervene.
• Time — the depth variable. Determines how far the ferment will go and what it will become. A ten-day sauerkraut and a six-month miso are not different recipes — they are different time horizons on the same family of reactions.

Chapter 1 of this Notebook does not introduce the variables. They are introduced here, in this Introduction, because they are the lens for everything that follows. Chapter 1 puts numbers on them, gives the ranges that work and the ranges that fail, and shows what happens when you push each variable past its range. From Chapter 1 onward, every chapter assumes you have those three variables in your head as a working frame.

The Safety Doctrine — Chapter 2 — sits between Chapter 1 and the foundation ferments because, having put the three variables in front of you, the next thing to do is name the situations where the variables can mislead you. Mold on the surface of an otherwise correct-looking ferment. A botulism risk in a low-acid, low-salt, oxygen-sealed jar. A nukazuke bed left untended for two weeks. A koji culture held above 35 °C. The Safety Doctrine is short. Its job is to be unambiguous about discard rules, so that the rest of the book can teach the cook to read a ferment without worrying that the reading itself might be the dangerous part.

After Chapter 2, the rest of the Notebook teaches by walking through six foundation ferments — sauerkraut, kimchi-style, nukazuke, shio-koji, vinegar pickles (as a boundary case, not actually a fermentation), and miso starter — and shows how the three variables and the discard rules apply to each. By the time you have read all six, the next ferment you encounter — anything from a kefir to a kombucha to a Japanese amazake — will be readable in the same frame. You will know which variable to look for. You will know what range that variable should be in. You will know what to discard.

That is the entire teaching of the book. The rest is detail.

A note on time

This is also the moment to say, out loud, that fermentation is the cooking system most demanding of the cook's patience and the most resistant to the cook's impatience. A sauerkraut takes ten to fourteen days. A nukazuke bed takes three weeks to mature and is then tended for years. A miso takes six months to two years. The Notebook can teach you which decisions matter at the start of the ferment. The Notebook cannot teach you to wait.

The cook who skims the Notebook looking for "the fastest fermentation" will find a chapter on vinegar pickles, which are not technically fermentation, and a chapter on shio-koji, which can be made in a week. Everything else operates on weeks-to-years. This is not a flaw in the book. It is the subject matter.

What the cook gets in exchange for the time is a flavor and a relationship to a substrate that no fresh cooking can produce. A bowl of soup made with two-year miso has, in its background, two years of enzymatic work. A radish dropped overnight into a healthy nukazuke bed has, in its skin, the accumulated character of every vegetable that has passed through that bed before it. A spoonful of homemade sauerkraut tastes, at the third week, of a chemistry that began the moment the cook salted the cabbage and chose the temperature of the room.

This is what the Notebook is, at its quietest, about. A way to enter that timescale on purpose. With the right numbers. Without the wrong guesses.

The rest of the book is the numbers.

From the Notebook — opening note to the cook

The single most useful instrument in fermentation is a kitchen scale accurate to one gram. The second most useful is a thermometer. The third is the calendar on the wall.

If you read no further than the next two chapters, those three instruments — and the discard rules in Chapter 2 — will protect you from the worst that fermentation can do. Everything after Chapter 3 is upside.

Three variables decide most of what happens in a fermentation jar. The rest of the variables — the substrate, the vessel, the cleanliness, the spice profile, the cook's mood — modulate the result. But salt percentage, temperature, and time decide which version of which ferment the jar produces, and whether it is the version the cook wanted.

This chapter sets the numbers on each. By the end of it, you should be able to look at a jar of brining cabbage and predict — within ranges — what it will be in three days, in ten days, and in three weeks. That predictive ability is what separates a cook who repeats their successes from a cook who repeats only their hope.

The chapter is in five sections. The first explains why the three variables form a system, not a checklist. The next three take each variable alone and put numbers on it. The fifth shows how the three interact when you change one without changing the others — the most common cause of fermentation failure in the home kitchen.

1.1 · One ecology, three knobs

The jar of cabbage on the counter is, at the microbial level, a small ecosystem. Hundreds of bacterial species are present in tiny numbers at the moment the jar is sealed: lactic acid bacteria (LAB) that the ferment depends on, mixed in with whatever else was on the cabbage and in the kitchen. What happens next is selection. The conditions inside the jar will favor some of those species, suppress others, and within seventy-two hours produce a population dominated — or not — by the LAB the cook wanted to win.

The conditions are set by three knobs the cook controls in advance.

• Salt percentage governs which organisms can compete. It is a chemical threshold. Above a certain salt concentration, most spoilage organisms cannot grow; below it, they can. The threshold is sharp, not gradual.
• Temperature governs how fast the competition runs. Bacterial metabolism roughly doubles every 10 °C above a baseline. The same ferment at 18 °C and at 28 °C is the same chemistry — but the 28 °C jar finishes in half the time, and the window where things can go wrong is also halved.
• Time governs how far the chemistry runs. A ferment stopped at day seven is a different food from the same ferment stopped at day twenty-one. The cook chooses which one by deciding when to move the jar to the refrigerator.

These are three knobs, not three independent dials. Change one and the others move with it. A higher temperature makes the ferment finish faster, which means the time horizon for the same flavor depth is shorter, which means the cook has less margin to read the jar before it goes past the target. A lower salt percentage opens the jar to organisms that work faster than the LAB, which means the temperature has to be lower to keep things in check, which means the cook has to commit to a cool corner of the kitchen for the whole window. The three knobs are coupled.

Most home fermentation failures, in our experience, are not because the cook picked one wrong variable. They are because the cook picked one variable correctly, and a second variable incorrectly, and the second variable pushed the first out of its safe range. The number-one example: a 2 % brine at 18 °C is reliably a sauerkraut. The same 2 % brine at 30 °C is sometimes a sauerkraut and sometimes a slimy disaster, because at 30 °C the LAB can no longer keep ahead of the other organisms. Cook chose right on salt, wrong on temperature, and the result was unpredictable.

The teaching frame this Notebook will keep returning to is: do not optimize any single variable. Match the three together to a known ferment, and only change one at a time when you experiment.

1.2 · Salt — the threshold variable

Salt is the variable the cook controls most precisely and the variable that, when set wrong, fails fastest. It is also the variable that most other fermentation books mishandle, by giving volumetric measurements that drift between brands of salt and by giving ranges so wide that the threshold meaning is lost.

The number is two percent.

One kilogram of shredded cabbage plus twenty grams of pure salt, mixed and pressed under its own brine, will reliably produce sauerkraut within ten to fourteen days at room temperature. This is the canonical sauerkraut ratio and, with slight variation, the canonical kimchi ratio (Korean kimchi typically runs 2.0–2.5 %, accounting for the salt drawn out by the pre-brining of cabbage). The German, Polish, and Alsatian sauerkraut traditions — developed independently over many centuries before anyone had a microscope — all converged on essentially the same percentage. They were not coordinating. They were tracking an ecological boundary that the cabbage itself enforces.

Why two percent

Salt does not sterilize the jar. It changes the competition. At the correct percentage, it favors the organisms this ferment depends on — the lactic acid bacteria, principally Leuconostoc mesenteroides in the first phase and Lactobacillus plantarum in the later phase — and disadvantages many of the organisms the cook does not want. The National Center for Home Food Preservation (NCHFP) frames this the same way: in fermented vegetables, salt is what allows the desired microbes to dominate the substrate before spoilage or pathogenic organisms can, and the recommended percentage should not be reduced for health reasons.

Two percent salt is the empirical percentage at which this competition consistently resolves in favor of the LAB in a home-kitchen setting at 18–22 °C. It is not a magic number; it is the ratio at which centuries of independent sauerkraut traditions converged. Lower percentages allow the wrong organisms to compete more effectively; higher percentages slow the LAB into a different kind of ferment (see §1.2 upper-bound discussion below).

What happens at the floor

Push the salt down to one percent and the window collapses. The brine becomes hospitable to organisms it should not be hospitable to. A 1 % ferment can still succeed if the cabbage is exceptionally clean, the temperature is cool, and the cook is lucky — but the failure mode, when it comes, is fast and unmistakable. The jar smells of sulfur or rotting meat within three days. There is no salvaging it. (See Chapter 2 for the discard rule.)

The risk does not scale linearly. The floor is a threshold, not a slope. This is one reason "salt-free wild fermentation" experiments produced by some enthusiasts are, from a food safety standpoint, an inadvisable approach: the difference between success and failure is the difference between a lucky jar and an unlucky one, with the cook unable to tell which one they have until the smell announces it.

What happens at the ceiling

Push the salt up past five percent and the opposite problem arrives. The LAB slow down. By 7–8 % they have nearly stopped, and the ferment "arrests" — it is no longer producing acid fast enough to reach the pH 4.0 mark that defines a safe and stable preserved food.

This is not always undesirable. Japanese long-aged tsukemono (nukazuke, takuan), traditional fish sauce production (Vietnamese nuoc mam, Roman garum, Japanese ishiru), and miso fermentation all use salt percentages between 5 % and 15 %. They are not trying to ferment quickly. They are trying to ferment slowly, over months or years, while letting enzymes — bacterial and fungal — do the work of breaking down proteins and developing flavor. Miso at 12 % salt is essentially a controlled, decades-tolerant ageing process, in which Aspergillus oryzae (koji) enzymes hydrolyze soy proteins into amino acids without the lactic acid bacteria ever gaining the upper hand.

These are different ferments with different goals. The salt percentage tells you which one you are making before you taste anything.

The math, the scale, the brutal simplicity

The math is brutally simple and absolutely unforgiving. The formula is salt mass divided by total mass (cabbage plus salt, or vegetable plus brine if using a wet brine), expressed as a percentage.

• One kilogram of cabbage plus 20 g of salt is 20 ÷ 1020 = 1.96 %. Close enough to call 2 %.
• One kilogram of cabbage plus 30 g of salt is 2.91 %. A noticeably saltier ferment that will run slower.
• One kilogram of cabbage plus 10 g of salt is 0.99 %. Entering the danger zone.

The difference between a beautiful jar of kraut and a slow disaster is, in this case, ten grams. There is no way to know that by feel. There is no way to know that by taste at the start, when the salt has barely begun to dissolve. The kitchen scale, accurate to one gram, is the only honest instrument in fermentation.

This is also why volume measurements — "a teaspoon of salt per pound of cabbage," the kind of instruction that appears in older American cookbooks — are unreliable for ferments. The bulk density of salt varies dramatically with grain size.

• A teaspoon of Diamond Crystal kosher salt weighs roughly 2.8 g.
• A teaspoon of Morton kosher salt weighs roughly 4.8 g.
• A teaspoon of fine table salt weighs around 6 g.

The same volumetric instruction can produce a 1 % ferment or a 2.5 % ferment depending on which box happens to be in the cupboard. For boiling pasta this does not matter. For determining which bacterial species inherits a kilogram of cabbage, it matters absolutely.

The practical rule

Weigh the vegetable. Weigh the salt. Calculate the ratio. Aim for the percentage that matches the ferment you want.

The cabbage does not care about your intentions. It cares about your numbers.

1.3 · Temperature — the speed variable

The second variable is temperature, and its job is to set the speed of the chemistry that the salt has decided to allow. Salt picks the players. Temperature picks the tempo.

The simplest rule

Most vegetable ferments run reliably at 18–22 °C. This is room temperature in a temperate-climate kitchen for most of the year. Within that window, lactic acid bacteria reproduce at a rate that out-paces the wrong organisms, the cook gets a working time window of about ten to fourteen days for a sauerkraut, and the ferment behaves predictably.

Push above 24 °C and things speed up. By 28 °C, the same sauerkraut may be finished in seven to ten days instead of fourteen — but the speed is also given to the organisms that the salt was supposed to suppress. The margin for the cook to read the jar shortens. A 28 °C ferment that smells off on day four cannot be rescued the way a 20 °C ferment that smells off on day eight can be. The window for intervention has collapsed.

Above 30 °C is where most home fermentation enters a high-risk zone. The LAB are still active, but slower than the competition. Spoilage organisms that the 2 % salt would normally outpace can now outpace the LAB if the temperature stays high. This is why summer fermentation in non-air-conditioned kitchens — especially in humid climates — is harder than winter fermentation, and why traditional fermentation cultures in hot regions (Korea, Japan, Vietnam) developed cellars, pits, and earthenware vessels designed to keep the ferment cooler than the room.

Push below 15 °C and the LAB slow down, but so does everything else. A cold ferment is generally safer than a hot one; it is just slower. A sauerkraut at 12 °C might take three weeks instead of two. A nukazuke bed kept in a 10 °C cellar takes months to mature. The cook trades patience for safety. This is the entire logic behind cellar-based traditional fermentation: cool, stable, slow, repeatable.

The numbers, ferment by ferment

The pattern: most ferments are happiest at human-comfortable kitchen temperature. The exceptions — shio-koji at 25–28 °C, miso at long-ambient — are noted because the cook will need to think about them differently.

The koji exception

A small but important note. Koji-based ferments (shio-koji, miso starter, sake mash) require Aspergillus oryzae enzyme activity, which has a specific temperature window — roughly 25–35 °C is its working range, with optimal enzyme production around 28–30 °C. Above 35 °C, the enzymes the koji produces denature. Above 40 °C, the koji culture itself stops being productive. Below 22 °C, the enzymes work too slowly to matter.

This means koji ferments are the one place in the home kitchen where temperature control is genuinely demanding. A shio-koji that sits on a counter at 18 °C for a week may not be a finished shio-koji; the enzymes have not had enough heat to do their work. A shio-koji that sits in a sun-warmed corner at 35 °C may have produced a different shio-koji than intended, with sharper flavor and less softening capacity.

The traditional Japanese solution is a koji-muro — a wood-walled fermentation room kept at constant 28–30 °C. The home-kitchen substitute is a yogurt maker or a sous-vide bath set to 28 °C. The cook who is unable to hold koji at 28 °C reliably should know that the ferment will work in a wider range, but the quality will vary, and a colder kitchen produces a milder, less developed shio-koji. (Chapter 3 gives the full shio-koji procedure with temperature handling.)

The temperature monitor

The cook does not need to measure constantly. A single check at the start of the ferment, ideally with a digital probe thermometer in the brine or alongside the jar, is enough to know whether the temperature window is being honored. After that, a rough sense of the kitchen — cool, warm, hot — is sufficient. The thermometer is most valuable in two situations: at the start of any ferment, and when something has gone wrong and the cook needs to retroactively diagnose which variable was outside its window.

1.4 · Time — the depth variable

The third variable is time, and it works differently from the first two. Salt and temperature are set at the beginning and largely held constant. Time is the variable the cook stops at a chosen point. The choice of when to stop is the choice of what the ferment will be.

The shape of fermentation time

In a generic vegetable ferment, the time course looks roughly like this:

• Day 0: cabbage shredded, salted, pressed under its own brine. Jar sealed (loosely — gases will need to escape). Internal pH around 6.2, neutral.
• Days 1–2: Leuconostoc mesenteroides — the first-phase LAB — multiplies rapidly in the brine. Gas production (CO₂) begins. The jar may swell slightly. Bubbles visible against the glass. Smell is fresh, lightly sour.
• Days 3–5: pH drops below 4.5. The lower pH starts to inhibit Leuconostoc itself, and Lactobacillus plantarum takes over. Acid production accelerates. Cabbage softens slightly. Smell becomes distinctly sour.
• Days 6–10: pH continues to drop, settling around 3.4–3.7 for a finished sauerkraut. Acid bite peaks. Cabbage texture stabilizes — softer than fresh, but still with structure. The ferment is approaching done.
• Days 11–14: refinement. The flavor sharpens, then mellows. The cook can taste daily from this point and decide when to call it done.
• Day 14 onward (if left at room temperature): ferment continues slowly. Beyond day 21, texture begins to break down. The cabbage loses crunch. Some cooks find this acceptable for cooked-into dishes; others move the jar to cold storage at day 10–14 to lock the texture in.

The same shape applies, with longer timescales, to other ferments. A kimchi runs through the same arc in 5–7 days, then is moved to cold storage where it continues to develop slowly for weeks. A miso runs through it in months, with the Aspergillus oryzae enzymes doing their hydrolysis work first and the LAB working in the background for the duration.

Why "done" is a decision, not an event

A ferment does not finish itself. It slows down. At some pH and acid load, the LAB themselves cannot reproduce further, and the chemistry effectively arrests. But this arrest is gradual. The cook decides "done" by choosing a moment to move the jar from room temperature to refrigerator (about 4 °C). The refrigerator drops the metabolic rate by roughly a factor of ten, and the ferment from that point develops only slowly, over weeks or months.

This is the central decision of fermentation: not how to start, but when to stop.

A cook who moves the sauerkraut to cold storage at day seven gets a milder, crunchier kraut. A cook who waits until day fourteen gets a sharper, softer kraut. Neither is wrong. Both are choices. The flavor of the resulting jar is determined more by the day-of-cold-storage decision than by almost anything that happened earlier (assuming the salt and temperature were correct).

Time horizons across ferments

The cook should hold the relevant timescale in mind before starting a ferment. A weekend project does not match miso. A six-month commitment does not match sauerkraut. The book the Notebook is trying to be is, in part, a frame for matching the cook's time horizon to the right ferment.

A note on patience

The most common time mistake in home fermentation is impatience: tasting on day three, deciding it does not taste like sauerkraut yet, and adding more salt or moving to fresh brine. The day-three jar is not a finished ferment. The day-three jar is Leuconostoc in mid-acidification. The cook who intervenes is not adjusting; they are interrupting the process. The right action on day three is to do nothing and let day seven happen.

This is why the calendar on the wall is the third most useful instrument (after scale and thermometer) in fermentation. Mark the start date. Mark the day-seven check date. Mark the day-fourteen done-or-continue date. Use those dates as the operational instructions. Read the jar on those days. Do not over-read on the days between.

1.5 · How the three interact (the diagnostic frame)

Most ferments succeed when the three variables match a known recipe. Most ferments fail when one variable is correct, a second variable is wrong, and the cook had not considered the interaction. This section is the diagnostic frame: when something has gone wrong, which variable was the cause?

The four most common failures and their causes

Failure 1: The ferment smells putrid, rotten, or of sulfur within 48 hours.

Cause, in order of likelihood:

1. Salt percentage too low. 1.5 % or below.
2. Temperature too high. Above 28 °C.
3. Cabbage exposed to air (not submerged under brine).

Discard. Do not save. Start over. Adjust salt to 2.0 %, move to a cooler location, ensure submersion. (See Chapter 2 for discard rule details.)

Failure 2: The ferment is "alive" — bubbling, fresh-smelling — but at day fourteen it still does not taste sour.

Cause:

1. Temperature too low (below 15 °C). The chemistry is running, just very slowly.
2. Salt percentage too high (above 5 %). The LAB are suppressed.

Action: give it more time. Move the jar to a slightly warmer corner (18–22 °C) and wait another seven to ten days. If the salt was actively high — the cook used 4 %+ — the ferment may be a slow-arrested ferment that is still safe but will take weeks longer to reach the target.

Failure 3: White film on the surface of the brine. Ferment otherwise smells fine.

Cause: kahm yeast. Common, mostly harmless, but indicates the ferment is at the edge of its safe window. Usually means temperature too high or oxygen contact (submersion failure).

Action: skim the surface gently. Re-submerge any floating solids. Move to a cooler location. If kahm returns, ferment is finishing late; consider moving to refrigerator earlier than planned to stop development.

Failure 4: Colored mold on the surface (green, black, blue, pink, white-fluffy).

Cause: aerobic contamination. Salt is insufficient, or submersion is insufficient, or temperature is too high, or all three.

Action: discard. This is the discard-first rule from Chapter 2. Mold of any color on the surface of a vegetable ferment is the signal to throw the jar away. The cook will be tempted to scrape and continue. The cook should not. The visible mold is the top of a mycelium that has likely already penetrated into the food.

The diagnostic question

When something feels off, the question is always:

Which of salt, temperature, time was outside its window?

If salt was correct (2.0 %) and temperature was correct (18–22 °C), time is almost never the cause of an off ferment in the first 14 days. The ferment may be slow, but it will not be unsafe. If salt was correct and temperature was wrong, the result is variable — sometimes fine, sometimes off, never reliable. If salt was wrong and temperature was wrong, the result is unsafe and should be discarded.

The cook who can answer this question for any given jar — which variable was outside its window? — has learned the diagnostic frame the rest of this Notebook builds on. Chapters 3 through 6 walk through specific ferments with this frame implicit. Chapter 2 — the next chapter — names the rules for when the diagnostic returns "throw it away" and when it returns "save it." Those rules are short, unambiguous, and the most important things in the book.

From the Notebook — Chapter 1 closing note

The three variables are the entire frame. Everything else in the book — the foundation ferments in Chapter 3, the failure modes in Chapter 5, the living relationships in Chapter 6 — is downstream of the three.

When you forget the rest, return to the three. When you encounter a new ferment, ask three questions: what salt percentage; what temperature window; what time horizon. When something has gone wrong, ask one question: which of the three was outside its window.

Reading the chapter once gives you the numbers. Reading the chapter three times — once a year for three years — builds the intuition that lets the cook do without the numbers.

This chapter is the load-bearing one. Every later chapter depends on it. The rules below are written to be short, unambiguous, and final. They are the discard rules. When you have read the rest of the book and need a refresher, return to this chapter first.

Why this chapter is here, and why it is short

Most cooking goes wrong in ways that are unpleasant and recoverable. A soup is over-salted. A sauce breaks. A loaf burns. None of these failures harm the cook beyond the inconvenience of starting over. Fermentation is different. Fermentation is the one cooking system where ignoring certain rules can produce food that hurts the person who eats it. The rules are not many — eight, in this chapter — but each one is non-negotiable. The chapter is short because length would dilute it. Each rule is one sentence, followed by the reason, followed by the operational test.

A rule without its reason is a rule that gets forgotten. A rule with its reason but no operational test is a rule the cook cannot apply at the moment of decision. This chapter gives all three for each rule.

What this Doctrine is, and what it is not

This Doctrine is about what to do when a ferment might be unsafe. It is not a health claim about fermented food when it is safe. The book makes no claim that fermented food benefits gut health, immunity, longevity, or any other physiological outcome. Safety, here, is about not getting sick. Flavor is the subject of every other chapter. Benefits are not the book's subject. If a fermentation source — book, podcast, social media — leads with health benefits and treats safety as an afterthought, treat the credibility of that source as questionable.

The Doctrine also draws three boundaries that the cook should hold in mind throughout the book:

• Vegetable lacto-fermentation (sauerkraut, kimchi, nukazuke) — in scope for this Notebook. The rules here apply.
• Vinegar pickles — not fermentation. The acid is added by the cook, not produced by microbes. These follow refrigeration-only rules and are discussed only as a boundary case in Chapter 3. They are not in the scope of "fermentation."
• Alcoholic fermentation (wine, beer, sake, kombucha booch) — different ecosystem, regulated differently in different jurisdictions, mostly out of scope for this book. Kombucha is the one borderline case (technically a mixed yeast-and-bacteria ferment with a tiny alcohol production) and is discussed with that caveat.

When the rest of the book says "ferment" without a qualifier, it means deliberate microbial fermentation in food — the lacto, koji, and mixed-culture systems Chapter 3 walks through. Vinegar pickles and alcohol production are named explicitly when they appear.

Rule 1 — Mold on the surface of a ferment is a discard signal, not a clean-up signal.

The rule. If you see colored mold (green, black, blue, pink, fluffy white-with-structure) on the surface of any home ferment, discard the entire jar. Do not scrape and continue. Do not eat what is underneath.

Why. Mold visible on the surface is the fruiting body of a mycelium that has been growing into the food for some time before it broke the surface. By the time the cook can see the mold, the hyphae of the same organism are likely already distributed throughout the soft food underneath. Some molds (notably Aspergillus flavus and Penicillium species) produce mycotoxins — aflatoxin, ochratoxin, patulin — that are stable to cooking, not killed by heat, and cumulatively harmful even in small doses. The cook cannot tell by looking which mold has produced toxin and which has not. The home test is: any visible mold means discard.

The exception that is not an exception. Kahm yeast — a flat, white-but-textureless film on the surface of an otherwise correct ferment — is not mold. It is a yeast film. It indicates the ferment is at the edge of its safe range (usually because of warm temperature or insufficient submersion) but is not in itself a discard signal. Skim it, re-submerge solids, move to a cooler location. If kahm yeast keeps returning, end the ferment by moving to cold storage. If colored mold appears on top of kahm yeast, the kahm reading was wrong; discard.

When in doubt, discard. The cost of discarding a correctly-fermenting jar that looked moldy is a kilogram of cabbage and a week of patience. The cost of eating a jar that looked like kahm yeast but was actually early mold is real harm. The asymmetry is so steep that the default action on any ambiguous surface film should be discard. The kahm exception in the paragraph above is for confident readers in clean kitchens; it is not the default action. The default action is discard.

Operational test. Ask: is the film flat and white with no structure, or does it have texture, fuzz, or color? Flat-white = kahm (recoverable if the rest of the ferment is correct). Fuzz or color = mold (discard). If you cannot decide which it is, the answer is discard.

Named structural exceptions — miso salt cap and nukazuke bed

Two foundation ferments have explicitly named, narrowly scoped exceptions to the surface-mold discard rule, because their structure or process provides a real physical or biological barrier between a surface contamination and the food the cook will eat. These exceptions are listed below because the book teaches both ferments; they are NOT invitations to extend the rule to any other ferment or to ambiguous cases.

Exception 1 — miso salt cap. A correctly-packed miso has a deliberate salt cap on the top surface. If colored mold appears on the salt cap and the cook can scrape the affected cap surface (1 cm deep) cleanly, and the paste below the salt cap is unaltered in smell and color, the paste below can be used. The 12 % salt content of the miso paste and the months-long ageing form the safety frame that the salt cap protects. If the smell or color of the paste below the salt cap is off, or the mold has penetrated below the cap, the entire tub is discarded. Ch.5 §5.2 names the operational test.

Exception 2 — nukazuke (rice-bran pickle bed). A nukazuke bed is a continuous-process ferment — the bed itself is the ferment, not the vegetables passing through. Surface mold can be skimmed (2–3 cm of the affected surface scooped off) and the bed below tasted; if the smell is clean and the bed has not visibly altered, the bed continues. If the smell is off, or the mold has penetrated below the surface, the entire bed is discarded. Ch.5 §5.14 names the operational test. Rule 7 (§SD-7) also governs the daily-stir relationship the bed requires.

These two exceptions are named because they reflect long-established culinary traditions where the food's structure or process provides a real barrier. They are not a softening of Rule 1 for the general case. For any other ferment, for any ferment where the exception's specific structural condition (intact salt cap, intact bed below the affected surface) does not hold, and for any ambiguity at the operational test, the default action remains discard the entire jar. The exceptions exist to acknowledge two centuries-old practices honestly, not to give the cook room to reason around the rule.

Order of evaluation for both named exceptions. Regardless of which exception applies, the cook must evaluate the ferment in this order: first structure (is the salt cap intact? is the bed still cohesive?), then color (is the material below the affected surface unaltered in color?), then smell (is the smell clean and characteristic of the ferment?), and only after all three visual and olfactory tests have clearly passed, then taste a small amount from a clean portion. Taste is never the first check. If any of structure, color, or smell fails, the exception does not apply and the ferment is discarded.

Rule 2 — Botulism risk is highest in low-acid, low-salt, anaerobic ferments.

The rule. Any sealed jar of fermenting food that combines low pH-shift, low salt, low oxygen, and room-temperature ageing carries botulism risk. This includes home-jarred low-acid vegetables, any oil infusion containing fresh garlic or fresh herbs, and any "fermented" jar that does not visibly bubble within 48 hours.

Why. Clostridium botulinum is a spore-forming bacterium that is harmless when oxygen is present but produces a potent toxin when the environment is anaerobic, low-acid (pH above 4.6), and warm. The toxin is heat-labile at boiling temperatures, but most home-fermented foods are eaten without re-boiling. The signal the cook normally relies on — a jar that smells off — is not reliable for botulism. The toxin is odorless and tasteless. The food may look and smell fine.

This is why a correctly-fermented sauerkraut (acidified by LAB to pH below 4.0, with 2 % salt) is safe — the acid is well below the Clostridium threshold. And why an oil infusion containing fresh garlic, fresh herbs, or low-acid vegetables at room temperature is dangerous — the oil is anaerobic, the inclusion is low-acid, and the conditions are exactly what Clostridium needs.

Operational test — by ferment family. The single-test framing "pH below 4.6 / salt above 2 % / bubbling by day 2" is a useful checklist for vegetable lacto-ferments but does not fit every ferment in this book. pH 4.6 is the regulatory boundary for acidified foods (FDA 21 CFR 114) and the Clostridium growth threshold; the Notebook uses it as a safety boundary explanation, not as a universal home-cooking test. Apply the family-appropriate check:

• Vegetable lacto-ferments (sauerkraut, kimchi, nukazuke) — use a tested salt percentage (§SD-8), keep solids submerged (§SD-5), and confirm active fermentation within the expected early window (bubbling by day 2 at 18–22 °C).
• Kombucha — salt is not the safety frame; adequate acidic starter liquid (at least 10 % of total volume) and a measurable pH drop below 4.0 by day 5 are. The starter is what suppresses Clostridium here, not salt.
• Koji and miso — salt percentage (11 %+ for shio-koji, 12 %+ for miso) and the correct temperature window are the safety frame. Bubbling is not expected and its absence is not a warning sign; the koji ferment is enzymatic.
• Oil infusions (garlic-in-oil, herb-in-oil, chili-in-oil) — not fermentation. Do not apply this test at all. Use refrigeration + 4-day rule, or do not make.

If any recipe does not match one of the four families above — especially oil-based or sealed-jar preparations that do not acidify — the recipe is not a safe home ferment. Use a tested recipe with a known acidifier, or do not ferment.

The aromatic-oil clarification

This is critical and worth its own paragraphs, because the marketing language around aromatic oils has become misleading.

Garlic-in-oil, herb-in-oil, chili-in-oil are not fermentations. They are flavored oils. The marketing word "fermented garlic oil" / "fermented herb oil" sometimes used by sellers and recipe writers refers to oil that has been infused; the garlic or herb has not undergone microbial fermentation. There are no significant LAB in the oil. There is no acidification. There is no bubbling. The cook who sees a recipe titled "fermented garlic oil" should treat it as a flavored oil with a safety risk, not as a fermentation that fits in this book.

For any aromatic oil prepared with fresh garlic or fresh herbs:

• Refrigerate below 4 °C from the moment of infusion.
• Use within four days.
• Do not store at room temperature.
• Do not gift such oils to others without naming the refrigeration requirement explicitly.

If you want a shelf-stable garlic oil, use dried garlic (a sufficiently low water activity for Clostridium not to grow) or commercially-produced garlic oil that has been treated to inhibit botulism (acidified, salted, or thermally processed under known parameters).

The known incident history of botulism from home-prepared garlic-in-oil — including hospitalizations and deaths — is the reason this rule is here. This rule also appears in the broader culinary safety doctrine for the site (see project_culinary_audit memory) and is reproduced here because fermentation readers are particularly likely to encounter recipes that mis-label oil infusions as "fermentation."

Rule 3 — Koji fermentation runs at 22–35 °C, which is not the vegetable-ferment window.

The rule. When growing or fermenting with Aspergillus oryzae (koji) — including shio-koji, miso starter, sake mash — the temperature must remain within 22 °C and 35 °C, with the optimum near 28–30 °C. Above 35 °C, koji enzyme production stalls and competing organisms (especially Bacillus) become a real risk. Above 40 °C, the koji culture itself is destroyed. Below 22 °C, enzyme activity is too slow to produce a meaningful ferment.

⚠️ Read this twice. Vegetable lacto-fermentation (sauerkraut, kimchi) works at 18–22 °C. Koji fermentation works at 22–35 °C. These are different windows. A sauerkraut held at 28 °C is at the upper edge of failure. A shio-koji held at 28 °C is in the middle of its optimum. The cook who mixes the two windows — keeping the sauerkraut warm because "fermentation likes warm," or keeping the shio-koji cool because "room temperature is fermentation" — is making different errors that fail in different ways. Hold each ferment to its own window. The Chapter 1 temperature table and the Chapter 3 walkthroughs name the window for every foundation ferment.

Why. Koji is unusual among fermentation organisms in being a mold (specifically Aspergillus oryzae) cultivated for its enzymes rather than for the LAB-style acid it does not produce. The enzymes — amylases (which break down starches) and proteases (which break down proteins) — work on the substrate at temperature. They are temperature-sensitive proteins. Outside the working window, they fold incorrectly and stop functioning, or denature outright.

A shio-koji that has spent its week at 18 °C is not a finished shio-koji; the enzymes have not had time at temperature to do their job. A shio-koji that has spent any time above 35 °C may have been overrun by Bacillus species, which produce off-flavors and, in rare cases, toxins. The cook does not get a clear visual signal for this; the jar can look normal and still be wrong.

Operational test. Use a thermometer at start of the ferment and at least once during the active window. If the kitchen cannot reliably hold 25–28 °C — for example, in winter or in a cold apartment — use a yogurt maker, a sous-vide bath set to 28 °C, or a turned-off oven with the light on (which warms the interior to roughly 28 °C). For miso, which takes months and runs at a wider temperature window, ambient room temperature is acceptable but should be checked seasonally; do not ferment miso in a kitchen that gets above 30 °C in summer.

Rule 4 — When you move a ferment to refrigeration, the chemistry effectively arrests.

The rule. Refrigeration at 0–4 °C reduces fermentation metabolism by roughly a factor of ten. Use this fact actively. Move a ferment to the refrigerator when it has reached the flavor you want; the jar will then develop only slowly, over weeks or months. This is the cook's stop button.

Why. This is not strictly a safety rule but a tool rule. The cook needs to know that "done" in fermentation is a decision the cook makes by moving the jar to cold storage. A sauerkraut left at room temperature past day fourteen will continue to develop — softer texture, sharper acid — and eventually break down. The same sauerkraut moved to the refrigerator at day ten is locked at that day's flavor and texture, give or take a slow refinement over the following weeks.

The safety side of this rule is the inverse. A ferment left at room temperature beyond its window can drift outside its safe range — kahm yeast can return, surface mold can appear, texture can break down. Refrigeration prevents these drifts. The cook who hesitates to refrigerate "because the ferment seems still active" should refrigerate anyway. Refrigerated ferments do not become unsafe; they just stop developing.

Operational test. Mark the start date on the jar. Mark a "decision date" — day 10 for sauerkraut, day 7 for kimchi, day 7 for shio-koji. On that date, taste the ferment, decide if it is at the flavor you want, and move to refrigerator. Do not wait for a clearer signal. The decision date is the signal.

Rule 5 — A fermenting vessel must keep solids under brine; air contact is failure waiting to happen.

The rule. Any vegetable ferment in brine must keep the vegetable submerged below the brine surface at all times. Floating solids contact air, and air contact produces mold, kahm yeast, or worse. Use a fermentation weight, a smaller jar inside the brine, a flat cabbage leaf as a barrier, or a vacuum-style fermentation lid. Do not ferment in a jar where solids float.

Why. The bacterial selection that fermentation depends on works only in anaerobic conditions in the brine. At the air interface, aerobic organisms — yeasts, molds, certain Pseudomonas species — can grow on the exposed vegetable surface even if the brine itself is fine. The cook who looks at a jar and sees half the cabbage above the brine is looking at a ferment that will fail at the surface even if the submerged portion is correct. The fail can be slow (kahm) or fast (mold), but it is unavoidable.

Operational test. After packing the ferment and pressing it down, the brine should cover all solids by at least 5 mm. If it does not, the cook must either add additional 2 % brine to bring the level up, or weight the solids down further. Check this on day one, day three, and day seven. If solids surface during the ferment (because of bubble buoyancy), push them back under or weight them.

Rule 6 — Cross-contamination between a failed ferment and the next ferment is a real risk.

The rule. A failed ferment — one that has been discarded because of mold, off-smell, or other failure — must be cleared, the jar and tools washed in hot soapy water, and the kitchen surface cleaned, before starting a new ferment in the same location. Wooden tools (chopsticks, spatulas) in particular should be sanitized or replaced.

Why. The organisms that caused the failure — mold spores, Pseudomonas, etc. — are now present in higher numbers in the immediate environment of the failed jar than they were before. A new ferment started without cleanup can be inoculated by those organisms from the start, biased against the LAB the cook is trying to favor. This is a quieter failure mode than the dramatic Rule 1 mold sighting, because the new ferment can look fine for several days before the inherited organisms catch up.

Operational test. After discarding a failed jar: hot soapy water on the jar (interior, exterior, lid), the fermentation weight, any tools that touched the failed ferment. Replace any wooden chopsticks or wooden spatulas if the failure involved colored mold. Wipe down the counter surface where the jar sat. Wait 24 hours before starting a new ferment in the same kitchen corner.

Rule 7 — A nukazuke bed, once started, is a living thing requiring daily attention.

The rule. A nukazuke bed is a long-term commitment. It must be stirred daily — by hand or with a wooden paddle — to maintain its bacterial balance and prevent surface mold or yeast bloom. A bed left unstirred for two weeks may not survive. A bed left unstirred for a month is almost certainly compromised and should be discarded.

Why. The nukazuke bed is unique among the foundation ferments in being a continuous-process fermentation. The bacterial population includes both LAB and certain wild yeasts. Daily stirring redistributes the bacteria, brings the surface back into the anaerobic interior, and prevents the formation of aerobic colonies on top. Skipped stirring lets the surface colonies dominate, which the bed cannot recover from once they have established.

For cooks who travel: a nukazuke bed can survive a week of refrigeration (covered, in the fridge, at 4 °C) without active attention, because the cold drops metabolism to near-zero. Anything longer than a week in cold storage stresses the bed. Anything longer than three days at room temperature without stirring is risk.

Operational test. Establish a stirring routine — morning or evening, paired with another habit (coffee, dinner prep) — so it is not forgotten. If the cook will be away for more than three days, move the bed to refrigeration. If the cook will be away for more than two weeks, give the bed to another cook to tend, or accept the likelihood that it will need to be restarted.

Rule 8 — Salt percentage below 1.5 % is a danger zone for vegetable ferments.

The rule. A vegetable ferment salted below 1.5 % by total weight should be treated as a failed ferment from the start. Do not try to "save" a low-salt ferment by adding salt mid-process. Discard, recalculate, and restart.

Why. This is the operational expression of Rule 2 (botulism risk) and of the salt-percentage discussion in Chapter 1. Below 1.5 % salt, the LAB cannot reliably outpace the wrong organisms. The cook will not know the ferment has gone wrong until day three or four, when the smell becomes obvious — but by then the brine is full of the wrong organisms, and adding salt at that point does not reset the population; it just inhibits everything, including any partial LAB recovery. The ferment is no longer a known ferment. Discard.

This rule is here because the most common version of this mistake is not deliberate. The cook intended 2 % but measured wrong, or used a salt brand with lower density than expected, or the kitchen scale was inaccurate. The 1.5 % floor is a safety margin: if a ferment was accidentally at 1.7 % salt, the cook can take the warning and move on; if it was at 1.0 %, the ferment is in actively dangerous territory and the rule says discard.

Operational test. Weigh the vegetable and the salt before mixing. Calculate the ratio (salt mass ÷ total mass × 100). If the result is below 1.5 %, add salt to bring it to 2.0 % before any fermentation begins. If you only realize on day two that you measured wrong, and the ferment is below 1.5 % salt, discard.

Quick-reference card — the eight rules

For printing and clipping near the kitchen scale.

Closing note — what this chapter does not cover

Four areas that are real but out of scope for v1:

1. Sourdough bread fermentation — different ecology, different safety profile, well-documented in baking books, not in this Notebook.
2. Alcoholic fermentation (wine, beer, sake) — regulated in many jurisdictions, requires a different safety frame.
3. Meat curing and charcuterie — botulism risk profile is sharper and requires curing salts (nitrites/nitrates) that home cooks should not improvise with. Out of scope here; refer to a charcuterie text.
4. Dairy fermentation (yogurt, kefir, cheese) — adjacent system, but with its own pasteurization considerations.

Future versions of the Notebook may add appendices for one or more of these. v1 holds the scope to vegetable ferments + koji + nuka + miso + vinegar pickle (boundary case), which is the scope the Atlas Ch.11 chapter established.

Six ferments are enough. The cook who learns these six can read almost any ferment they encounter afterward, because every other ferment in the world is some variation on the variables these six lay out. The chapter assumes Chapter 1 (Salt, Time, Temperature) and Chapter 2 (Safety Doctrine) are in the reader's hand. Each ferment cites the §SD-1 through §SD-8 rules verbatim — when this chapter says "if mold appears, see §SD-1," it means the reader can either remember the rule or turn back to Chapter 2. Both work.

How this chapter is organized

Each of the six ferments below uses the same eight-section structure:

1. What it is — the ferment in one paragraph, including which Atlas Ch.11 axis it primarily teaches.
2. Ratio / grams — the numerical specification. Salt percentage, water mass, substrate mass, in grams. Volumetric measurements (cups, teaspoons) do not appear.
3. Temperature window — the °C range that works, and the failure mode above and below.
4. Time window — active fermentation duration and decision-date guidance.
5. Daily observation — what to look for on day 1, day 3, day 7, day 14 (where applicable). The cook's eyes are the instrument; this section calibrates them.
6. Failure modes — three to five named ways the ferment goes wrong, each with cause and the §SD rule that applies.
7. Recovery / discard rule — when the ferment is salvageable, when it is not. The default for any ambiguity is discard (§SD-1).
8. Boundary notes — what this ferment is not, and what nearby practices it should not be confused with.

The six are in pedagogical order:

1. Sauerkraut — basics: salt, water, anaerobic, mold judgment.
2. Shio-koji — the koji temperature exception (§SD-3), safely.
3. Miso-style bean paste — long ageing and surface management.
4. Vinegar pickle — the boundary that explains "this is not fermentation."
5. Kombucha — sugar, acid, and the alcohol-boundary question.
6. Nukazuke — the living relationship, and the daily-discard judgment (§SD-7).

A cook who works through all six in this order learns the system. After this chapter, encountering any other ferment — kefir, doenjang, fish sauce, lassi, gari, sourdough culture — becomes a matter of mapping it onto the variables already in hand.

§3.1 — Sauerkraut (cabbage lacto-fermentation)

What it is

Sauerkraut is the teaching ferment. It is the simplest combination of substrate (shredded cabbage), salt (2 % by weight), and the cook's hand pressing the cabbage under its own brine. The cabbage releases water; the salt sets the brine; the lactic acid bacteria (LAB) already present on the cabbage leaves multiply under the brine; the brine acidifies; the result is sauerkraut. It teaches every axis in the book: salt percentage (§SD-8), submersion (§SD-5), the decision date (§SD-4), the mold rule (§SD-1), and the discard threshold (§SD-1, §SD-8). Master sauerkraut and most other vegetable lacto-ferments — kimchi, curtido, giardiniera-style fermented relishes — are variations on its variables.

Atlas axis: microbial + salt.

Ratio / grams

For one standard sauerkraut batch:

Larger or smaller batches scale linearly. For 500 g of cabbage, use 10 g of salt. For 2,000 g of cabbage, use 40 g of salt. Always 2.0 %. Never below 1.5 % (§SD-8).

Temperature window

Working: 18–22 °C. Higher temperatures speed the chemistry but also speed the wrong organisms; below 28 °C the ferment is still safe but at 28+ °C the failure modes (sulfur smell, slime, surface mold) become much more likely.

Lower: down to 12 °C is acceptable. The ferment will run more slowly — 18–21 days instead of 10–14 — but the result is reliable and often cleaner-tasting than a warmer ferment. Traditional German cellar fermentation is exactly this.

Upper: above 28 °C is outside this Notebook's recommended home-kitchen window and is a high-risk zone for vegetable ferments — the LAB no longer reliably outpace the wrong organisms, and failure modes (sulfur smell, slime, surface mold) become common. Do not start the ferment there unless the cook can actively control temperature. Move the jar to a cooler location (refrigerator if necessary) or do not start the ferment in summer if the kitchen cannot stay below 28 °C.

Time window

Day 0 — pack, salt, press, weight, seal loosely. Days 10–14 — taste test; if the brine is sour, the cabbage tender but still crunchy, and there is no off-smell, the ferment is at the decision point (§SD-4). Decision date — day 10 for a milder, crunchier sauerkraut; day 14 for a sharper, softer one. Move to refrigeration on that date. Cold storage — the kraut continues to develop slowly for months. Shelf life in refrigeration is 6+ months for a correctly-made sauerkraut.

Daily observation

• Day 1 — bubbles begin forming against the glass. Brine level rises slightly. Smell is fresh, faintly cabbage-and-salt.
• Day 3 — bubbling pronounced. Brine has cloudy appearance. Smell is distinctly lactic-sour. pH (if measured) is below 4.5.
• Day 7 — bubbling slowing. Cabbage texture has softened slightly but remains structural. Brine is fully cloudy. Smell is sharp, sauerkraut-recognizable. pH is around 3.6–3.8.
• Day 10–14 — bubbling minimal. Flavor is balanced sour. This is the decision date.
• Day 21 and beyond — if left at room temperature, cabbage texture begins breaking down. Move to refrigeration before this point unless the cook specifically wants a soft, well-acidified kraut for cooking-in.

Failure modes

Recovery / discard rule

The default action on any ambiguous reading is discard (§SD-1). Sauerkraut is the cheapest ferment to throw away — a kilogram of cabbage and ten days of patience — so the cook should err on the discard side without hesitation. The expensive mistake in sauerkraut is eating a jar that should have been discarded, not throwing away a jar that might have been fine.

Specifically: any visible mold, any sulfur or putrid smell, any pink or slimy texture, any salt-below-1.5 % calculation — discard, clean the vessel (§SD-6), start over.

Boundary notes

• Sauerkraut is not kimchi. Kimchi adds aromatics (gochugaru, garlic, ginger, fish sauce) on top of the same cabbage-LAB-salt mechanism. The variables are the same; the flavor layering differs. The Notebook covers kimchi as a variation in Chapter 4 (Reading the Ferment).
• Sauerkraut is not coleslaw. A coleslaw uses raw cabbage with vinegar dressing — no LAB, no acidification by microbes. Sauerkraut and coleslaw look similar in the bowl and are unrelated in the chemistry.
• Sauerkraut is not "fermented salad." The phrase sometimes appears in casual writing. There is no such category. A vegetable is either fermenting (the LAB is acidifying the brine) or it is not. Sauerkraut is fermenting; salad is not.

§3.2 — Shio-koji (rice koji marinade)

What it is

Shio-koji is the simplest koji ferment and the cleanest way to teach the koji-temperature exception (§SD-3). It is cooked rice that has been inoculated with Aspergillus oryzae (koji-kin) and grown into a fluffy white culture, combined with salt and water, and left at 25–28 °C for about a week so the koji enzymes — amylases and proteases — hydrolyze the rice starches into sugars and produce a sweet-savory liquid marinade. Shio-koji teaches a different ecosystem from sauerkraut: there is no LAB, no acidification, no bubbling. The chemistry is enzymatic, not microbial-acid. This makes the safety frame different — and naming that difference is the entire reason shio-koji appears here second.

Atlas axis: microbial (enzymatic) + salt.

Ratio / grams

For one standard shio-koji batch:

Total mass = 540 g. Salt as % of total = 60 ÷ 540 = 11.1 % — significantly higher than vegetable ferments. The high salt is correct for koji ferments; do not reduce it to "make it healthier." The high salt is what makes the long ferment safe (§SD-2).

Temperature window

Working: 22–35 °C. Optimum near 28–30 °C. This is different from the vegetable-ferment window (§SD-3). A shio-koji held at 18 °C will not finish; the enzymes work too slowly. A shio-koji held above 35 °C is at risk of Bacillus contamination, off-flavor development, and (rarely) toxin production.

The practical solution: a yogurt maker, a sous-vide bath set to 28 °C, or a turned-off oven with the light on (which warms the interior to roughly 28 °C). If none of these is available, place the jar in the warmest stable corner of the kitchen and use a thermometer at the start and again at day 3.

Time window

Day 0 — mix koji + salt + water; cover loosely; place at 25–28 °C. Days 1–3 — stir daily with a clean spoon (to redistribute the enzymes and avoid surface drying). Day 5 — taste the liquid. It should be sweet, salty, slightly fruity. If it is still tasting raw or grainy, continue. Day 7–10 — finished. The koji grains have softened; the liquid is opaque and slightly viscous; the smell is faintly sweet-sour like sake lees. Decision date — day 7 for a cleaner, milder shio-koji; day 10 for a richer, more enzymatic one. Move to refrigeration. Shelf life refrigerated is 3 months at standard freshness; longer if the cook accepts gradual flavor development.

Daily observation

• Day 1 — no visible change. Smell is fresh rice + salt.
• Day 3 — the rice grains are softening; the liquid is starting to opaque. Smell is faintly sweet.
• Day 5 — the liquid has thickened; rice grains have broken down. Smell is distinctly sweet-sour.
• Day 7 — the liquid is opaque, slightly viscous; the grains have mostly dissolved. Smell is sweet, faintly fruity. This is the decision date.
• Day 14+ — if held at temperature past day 10, the liquid darkens (Maillard browning); the flavor becomes deeper but the texture is the same.

Failure modes

Recovery / discard rule

The shio-koji failures are mostly discoloration or off-smell. The cook does not "save" a shio-koji that has gone wrong; the salt content is too high to dilute meaningfully without ruining the ferment, and the enzyme balance cannot be reset. Discard, clean (§SD-6), restart with fresh koji.

Boundary notes

• Shio-koji is not miso. Both use Aspergillus oryzae, but miso uses soybeans (or other beans) as the substrate and runs for months to years. Shio-koji uses rice koji + salt + water and runs for a week.
• Shio-koji is not amazake. Amazake is sweet (no salt) and ferments at higher temperatures (50–55 °C) for a shorter time. Shio-koji is salty and runs at moderate temperatures.
• Shio-koji is not a "fermented seasoning" in the marketing sense. It is a specific Japanese enzymatic ferment with known parameters. Recipes labeled "fermented seasoning" online sometimes mean shio-koji and sometimes mean something else; the cook should verify.

§3.3 — Miso-style bean paste (small-batch demonstration)

What it is

Miso is the patient ferment. It teaches what time does at the timescale of months and years — what no shortcut can substitute for. A small batch of miso, started with a cook's hand pressing soybeans and salt and koji into a tub, will spend six months to two years quietly hydrolyzing into a paste that did not exist when it began. This section describes a small-batch demonstration — about 1.5 kg of finished miso — sized so that a home cook can keep a single tub on a shelf in a cool corner of the kitchen and check it monthly. Industrial miso is a different scale and uses different controls; this Notebook teaches the home version.

Atlas axis: microbial (enzymatic) + salt + patient.

Ratio / grams

For one small-batch miso:

The paste is mashed (food processor, mortar, or by hand) into a coarse-grained consistency. Pack into a fermentation tub or large jar. Cover with a layer of salt on the top surface (the "salt cap") to inhibit surface mold. Weight the surface to keep the paste dense and oxygen-free underneath.

Temperature window

Working: 15–25 °C ambient. Miso is one of the few ferments that runs across seasons. Summer fermentation is faster (heavier flavor development), winter is slower (cleaner flavor). Above 30 °C is risk territory — off-flavor development becomes likely.

Lower: down to 10 °C (a cool pantry) — works, but the ferment will take 12+ months to mature.

Upper: do not let the kitchen exceed 30 °C for extended periods. In hot summers, move the tub to a cool basement, pantry, or temperature-controlled corner.

Time window

Day 0 — month 0 — pack, salt-cap, weight, seal. Month 1 — check: surface should look unchanged. The salt cap should be firm and white. No colored mold (§SD-1). Month 2–3 — slight liquid (tamari) may pool on top. This is the soy proteins liquefying; it is correct. Month 6 — first usable miso. Sweet-mellow flavor. Suitable for soups. Month 12 — deeper, more complex. Suitable for marinades and condiments. Month 18–24 — fully developed. Dark, savory, deeply aged.

The cook chooses when to call it finished by tasting at the 6-, 12-, and 18-month marks and deciding which depth they want.

Daily observation

Miso does not need daily attention. It needs monthly check-ins.

• Month 1 check — surface still salt-capped and firm. No colored mold.
• Month 3 check — surface may have some surface yeast or kahm-style film (§SD-1 kahm exception). Some surface darkening is normal.
• Month 6 check — paste is now sufficiently developed that the cook can scrape off the top 5 mm (any surface yeast or oxidized paste) and taste from underneath.
• Annual checks — monitor for §SD-1 mold; otherwise leave alone.

Failure modes

Recovery / discard rule

Miso is the one ferment where the salt-cap structure provides a partial defense against surface mold (§SD-1) — the cook can scrape and evaluate the paste below. First evaluate structure, color, and smell. Taste only after the named structural exception has clearly passed its visual and smell tests. If the paste below shows discoloration, smells off, or has visible mold below the salt cap, discard the entire tub. Do not eat miso that has lost its surface seal entirely. The 12 % salt + long ageing is what makes miso safe; if that integrity has been broken, the safety is broken too.

Boundary notes

• This is a home small-batch demonstration, not commercial miso. Commercial miso is fermented under controlled conditions in dedicated facilities. Home miso is a relationship the cook tends, and the variability is real.
• Miso is not shio-koji. Shio-koji is rice-only and weeks; miso is bean-and-koji and months-to-years.
• Miso is not doenjang. Korean doenjang is a similar concept (fermented bean paste) but the strain, salt percentage, and ageing pattern differ. This Notebook does not teach doenjang in v1.
• Miso is not "fermented soy." Soy sauce, natto, tempeh, douchi are all different ferments of soybeans with different organisms and methods. Naming each correctly matters.

§3.4 — Vinegar pickle (the boundary case)

What it is

This is the section that exists to teach what fermentation is by naming what it is not. Vinegar pickles — cucumbers in vinegar brine, pickled red onions, gari (Japanese sushi ginger), most American "refrigerator pickles" — are not fermentations. The acid in them is added by the cook in the form of vinegar (acetic acid). No microbial fermentation occurs. The vegetable is preserved by the low pH the vinegar provides, but the LAB never get involved, the chemistry never has the time-and-temperature shape of a sauerkraut, and the safety frame is therefore different.

Vinegar pickles belong in the refrigerator. They are not pantry food. The shelf life is two to three weeks for most home preparations, longer for commercial canned versions (which are heat-processed). This section is here so the reader does not confuse the two categories.

Atlas axis: not in the four fermentation axes. Acid-only preservation, no microbial component.

Ratio / grams

For a standard refrigerator vinegar pickle (e.g., quick cucumber pickle):

This is not a fermentation specification. There is no LAB. The vegetable is submerged in the vinegar-water-salt-sugar brine and refrigerated immediately.

Temperature window

Refrigerated, 0–4 °C, from the moment of preparation. Do not leave a vinegar pickle at room temperature. The acid alone is not sufficient to prevent spoilage at warm temperatures unless the jar has been heat-processed and sealed (a canning operation outside the scope of this book).

Time window

24 hours minimum for the flavor to develop. Eat within 2–3 weeks of preparation.

Daily observation

Almost nothing to observe day-to-day. The pickle does not change visibly in the refrigerator. The cook tastes one slice at the 24-hour mark to confirm flavor development.

Failure modes

Recovery / discard rule

For vinegar pickles, the discard threshold is the same as §SD-1: any visible mold, any off-smell, any sliminess means discard. Do not try to rinse and save. The vinegar pickle is the cheapest ferment-adjacent preparation to make; throw it away and start over.

Boundary notes

• Vinegar pickle is not sauerkraut, kimchi, nor any vegetable lacto-ferment. The cabbage in vinegar is a vinegar pickle. The cabbage in salt brine is a sauerkraut. The two are different food categories.
• Vinegar pickle is not "instant fermented pickle." That phrase, used in some marketing, is contradictory. Either it is vinegar (no fermentation, eat fresh) or it is fermented (LAB, requires time and the discard rules of §SD-1 through §SD-8).
• Vinegar pickle is not home canning. Home canning is a heat-process operation with its own safety rules (USDA, NCHFP guidelines) outside the scope of this book.
• Vinegar pickle is not safe at room temperature. This is the most important boundary note in this section. A jar of pickled onions on the counter for a week is a different category of risk from a refrigerated one. Refrigerate from the start.

§3.5 — Kombucha

What it is

Kombucha teaches the sugar-acid-alcohol boundary. It is a mixed-culture ferment of sweetened tea — black or green tea steeped with sugar — inoculated with a "SCOBY" (Symbiotic Culture Of Bacteria and Yeast), which is a cellulose mat hosting Acetobacter bacteria and various yeast strains. The yeasts ferment the sugar into alcohol; the bacteria oxidize the alcohol into acetic acid. The result, after roughly 7–14 days, is a slightly carbonated, sweet-sour tea with about 0.5–2 % residual alcohol. Kombucha is a borderline case in this book: it is microbial fermentation (in scope), but it produces alcohol (regulated differently in different jurisdictions), and the alcohol level shifts depending on ferment time and temperature.

Atlas axis: microbial + acid + (small) alcohol production. This is the only ferment in this book that produces alcohol in any meaningful quantity.

Ratio / grams

For one standard kombucha batch (1 L):

The salt content is zero. The kombucha's safety comes from the acidity it develops (pH dropping below 4.0 within 48 hours) and the starter liquid that maintains acidity from day zero.

Temperature window

Working: 22–26 °C. Above 28 °C the SCOBY is stressed and bacterial-yeast balance shifts toward yeast (more alcohol, less acid). Below 18 °C the ferment is too slow and the starter acidity may not hold against competing organisms.

Time window

Day 0 — sweetened tea + starter + SCOBY in jar, covered with a cloth (not sealed; the ferment is aerobic at the surface). Day 3 — taste; should be slightly sweet, slightly tart. Day 7 — taste; sweetness gone, sourness developed. This is the decision date for a milder kombucha. Day 10–14 — full sourness, more acetic-acid character. Decision date for a sharper kombucha. Day 14+ — alcohol may rise; vinegar character becomes dominant; bottle and refrigerate before this point.

After the primary ferment, the kombucha is bottled (with optional fruit or flavorings for a secondary ferment) and refrigerated. Refrigeration arrests further development (§SD-4).

Daily observation

• Day 1 — SCOBY may sink or float; new thin layer forming on the surface. Tea is still tea-colored.
• Day 3 — tea is paler, slightly tart on the tongue. Fizzing if poured.
• Day 5 — pH below 4.0 (if measured). New SCOBY layer forming on top.
• Day 7 — recognizably kombucha; sweet-sour balance.
• Day 10+ — sourness dominates; new SCOBY layer 5–10 mm thick.

Failure modes

Recovery / discard rule

For kombucha, the discard threshold is §SD-1 mold and §SD-2 botulism-adjacent risks. The visual ambiguity of a "healthy" SCOBY (which can look strange to a new cook) means the first-time fermenter should be especially conservative — when in doubt, discard the batch and the SCOBY, and start with a fresh SCOBY from a reputable source.

Boundary notes

• Kombucha is not strictly "tea." It is fermented sweet tea; the chemistry has changed it materially.
• Kombucha is not "alcohol-free." Home-fermented kombucha typically contains 0.5–2 % alcohol; commercial kombucha is regulated differently in different jurisdictions (in the US, must be under 0.5 % to be sold as a non-alcoholic beverage). The Notebook does not advise on regulatory compliance for selling.
• Kombucha is not a soda alternative for children or pregnant readers. The alcohol content, while small, is real. The Notebook makes no recommendation about who should or should not drink kombucha; it states the chemistry.
• Kombucha is not lacto-fermentation. The dominant acid is acetic, not lactic. The organisms are Acetobacter + yeast, not LAB. The pH frame (§SD-2 botulism threshold) still applies, but the route there is different.

§3.6 — Nukazuke (rice-bran pickle bed)

What it is

Nukazuke is the ferment that teaches the living relationship (§SD-7). A nukazuke bed is a bed of fermented rice bran in which vegetables are buried overnight or for a day, pickled by the bed's resident LAB and salt content, and removed for eating. The bed itself is the ferment; the vegetables passing through are not the ferment — they are tasting the chemistry of the bed. A new bed takes about three weeks to mature and is then tended for years. A nukazuke bed in a Japanese household may pass between generations. The discipline the bed demands — daily stirring, attention to surface, refrigeration during travel — is the heart of what this section teaches.

Atlas axis: microbial + salt + relationship.

Ratio / grams

For establishing a small nukazuke bed:

The bed lives in a ceramic, glass, or wooden tub. The cook stirs it daily by hand (gloves OK or bare hand; the bed has been historically tended by hand to allow microbial transfer from the cook's skin — a tradition, not a sterility argument).

Temperature window

Working: 15–22 °C ambient. Above 28 °C the bed is at risk of yeast overgrowth and surface yeast bloom. Below 12 °C the bed slows down; refrigerator overnight storage is OK but the bed should return to room temperature for active fermentation.

A summer kitchen above 30 °C is dangerous to a nukazuke bed; the bed should be refrigerated continuously during heat waves (which slows everything but keeps the bed alive).

Time window

Establishment phase (weeks 1–3) — daily stirring; sacrificial vegetables buried for 24-hour cycles and discarded; the bed develops its target population. Mature phase (week 4 onward) — the bed is ready to pickle vegetables for eating. Burial time varies by vegetable: cucumber 4–8 hours, daikon 12–24 hours, cabbage 24–48 hours. Long-term (months and years) — the bed is replenished by adding small amounts of fresh nuka + salt as it depletes; the bed is stirred daily; the bed is tasted weekly to monitor its character; the bed continues indefinitely.

Daily observation

The nukazuke bed requires daily attention. This is what makes it the §SD-7 ferment.

• Daily — stir the bed thoroughly by hand or with a wooden paddle. Bring the surface into the interior. Smell the bed: it should smell pleasantly of fermented rice and salt, with a slight earthy character.
• Weekly — taste a vegetable from the bed (a quick-pickle test). Evaluate salinity, sourness, depth.
• Monthly — add a small replenishment of fresh nuka + salt (about 5–10 % of bed mass) if the bed has thinned from vegetable handling.
• Annually (or as needed) — replace 20–30 % of the bed with fresh nuka if the bed has become too acidic, too watery, or too fatigued.

Failure modes

Recovery / discard rule

The nukazuke bed has its own §SD-7 rule, but inside §SD-7 the §SD-1 mold rule still applies. The cook's daily relationship with the bed is what catches problems early; a bed checked daily can usually have minor surface yeast skimmed without alarm. A bed left unstirred for two weeks should be assessed carefully and, if there is any doubt, discarded. The default action on a bed of unknown state is to start fresh.

Boundary notes

• Nukazuke is not "Japanese sauerkraut." Both are LAB ferments, but the nukazuke bed is a continuous-process ferment with a permanent ecosystem; sauerkraut is a single-batch ferment that finishes and is moved to refrigeration.
• Nukazuke is not commercially-bagged-pickle. Commercial Japanese pickles labeled nukazuke may be vinegar-flavored quick-pickles that imitate the flavor without the bed; reading the label matters.
• Nukazuke does not work with all vegetables. The traditional set is cucumber, daikon, cabbage, eggplant, carrot, kabu, and a few others. Watery vegetables (tomato) and very dense ones (raw potato) do not pickle well.
• The cook's hand in the bed is tradition. The historical practice of bare-hand stirring is part of the relationship-with-the-bed framing; it is not a sterility recommendation. The Notebook neither endorses nor forbids hand-stirring; it names the tradition. Use gloves if preferred.

Closing — what the six ferments teach together

If the cook has read the six sections above, they have encountered:

• The teaching ferment (sauerkraut) — every variable, smallest stakes.
• The koji exception (shio-koji) — different temperature window, different ecosystem, named explicitly so it is not confused with vegetable ferments.
• The patient ferment (miso) — what time at the timescale of months does, and what surface management means at long horizons.
• The boundary case (vinegar pickle) — what fermentation is not, named here to prevent confusion in the rest of the cook's reading.
• The alcohol-adjacent case (kombucha) — the one ferment in this book where alcohol production matters, and the regulatory frame is named without prescription.
• The living relationship (nukazuke) — the one ferment that requires daily attention and teaches the cook to develop intuition by reading a bed every day.

Chapter 4 of the Notebook walks through the day-3, day-7, day-14 reading frame applied across these six ferments side by side. Chapter 5 catalogs the 24 documented failure modes with their structural causes and the §SD-1 through §SD-8 rules that apply. Chapter 6 returns to the nukazuke bed and the kombucha SCOBY as the living-relationship case studies — what it means to tend an ecosystem the cook started.

The cook who has finished Chapter 3 has the operational vocabulary to read almost any home ferment they will encounter. The remaining chapters add depth; the foundation is here.

From the Notebook — Chapter 3 closing note

Three things to keep in mind across all six ferments:

1. The §SD-1 discard rule applies to all six. When in doubt, discard.
2. The salt percentages are not interchangeable. A 2 % sauerkraut is correct; a 2 % miso is dangerous (§SD-2). Hold each ferment to its own number.
3. The temperature windows are not interchangeable. A 28 °C sauerkraut is at the upper failure edge; a 28 °C shio-koji is at the optimum (§SD-3). Hold each to its own window.

The book repeats these three things because they are the three things the cook is most likely to forget.

Chapter 3 was about what to make. Chapter 4 is about how to look. The six foundation ferments from Chapter 3 are all observable, day by day, through the same six channels: smell, surface, texture, brine, temperature, time. This chapter calibrates those channels. By the end of it, the cook should be able to walk past a fermenting jar on the counter and know — in a few seconds — whether the jar is on track, off track but recoverable, or at the §SD-1 discard threshold. The remaining chapters of the Notebook depend on this skill. The Failure Modes catalog in Chapter 5 is, in part, a structured tour of what the trained eye sees and decides.

§4.1 — The ferment is not a timer

Recipes give the cook a number of days. The ferment does not know about the recipe. A sauerkraut at 12 °C and a sauerkraut at 22 °C are running the same chemistry at different speeds; the 12 °C jar at day 14 is at the same stage as the 22 °C jar at day 7. A shio-koji at 28 °C reaches its flavor at day 7; the same shio-koji at 22 °C may need day 10 or 12. The day-count on the label is a useful approximation. It is not the instruction.

The instruction is observation. The cook decides "done" by looking at the jar, smelling it, and tasting it — not by reading the date on the lid. The calendar is the third most useful instrument in fermentation (after the scale and the thermometer), but it is not the authority. The authority is the ferment, and the ferment speaks through the channels this chapter will name.

This is also why the §SD-4 decision-date rule says "decision date," not "completion date." A ferment does not finish on its own. It slows down. The cook chooses when to refrigerate, and that choice is made by reading the jar, not by counting days.

The chapter's central claim: almost every fermentation success and almost every fermentation failure can be observed before it is irreversible. The cook who learns to read the channels can intervene before damage; the cook who only reads the calendar finds out at day fourteen what should have been clear at day three.

§4.2 — Smell

The nose is the first instrument the cook calibrates, because smell precedes texture, color, and bubbling as a signal. A ferment that smells right almost always is right; a ferment that smells wrong almost always is wrong. Smell is also the channel where the cook's instinct sharpens fastest — by the third or fourth ferment, the cook recognizes correctness without thinking about it.

The categories the cook should learn to distinguish:

Clean lactic sour — sauerkraut at day 7, kimchi at day 5, nukazuke pickle freshly removed from the bed. The character is bright, slightly tart, fresh, similar to good yogurt or buttermilk. The smell announces "LAB are working." This is the target smell for any vegetable lacto-ferment.

Sweet-sour-fruity (koji-enzymatic) — shio-koji at day 5, miso at month 1. The character is sweet at the front, slightly fermenty-fruity behind, faintly reminiscent of sake lees. No sharp acid. This is the koji ferment's correct smell; it is not the sauerkraut smell, and a sauerkraut that smells this way is in a different ferment than the cook intended (probably temperature or salt error).

Mild alcohol — kombucha at day 5 onwards. A faint, slightly sweet ethanol note, like the smell of low-alcohol beer or a wine glass left out for an hour. This is correct for kombucha (§3.5 boundary case) and should not appear in a vegetable lacto-ferment. If a sauerkraut smells alcoholic, something has gone wrong — probably yeast overgrowth or sugar contamination.

Bread-yeast (kahm) — a vegetable ferment surface developing kahm yeast smells slightly bready, with a faintly sweet-musty undertone. This is the kahm exception of §SD-1 — skim the surface, re-submerge, move to a cooler location.

Solvent / acetone / nail polish — any ferment with this smell is in trouble. The chemistry is wrong; the wrong yeasts or Bacillus species are dominant. This smell appears in low-salt sauerkrauts, in over-warm kombuchas, and in shio-koji that has spent time above 35 °C (§SD-3). Action: discard.

Sulfur, rotten meat, ammonia — these are the §SD-8 and §SD-1 smells. They appear in low-salt vegetable ferments within 48 hours, or in miso that has been kept too warm too long. There is no salvage. Discard.

Putrid in any form — if the cook cannot place the smell but it raises an alarm, the alarm is the data. Discard. The cost of a discarded jar is a kilogram of cabbage and a week of patience; the cost of trusting an alarming smell is real harm.

The cook's nose calibrates over weeks of practice. The first ferment is uncertain; by the fifth, the cook knows. The Notebook recommends the cook smell every jar at every observation date, even when the jar is clearly fine, because the calibration is built by the daily comparison.

§4.3 — Surface

The surface of a ferment is the most visually informative channel. The cook looks at the jar from above (lid off, briefly) and from the side (through the glass) and reads the surface state.

Bubbles against the glass and rising through the brine — day 1 to day 5 in any active vegetable lacto-ferment. The bubbles are CO₂ from LAB metabolism. Vigorous bubbling on day 2 is correct; absent bubbling on day 3 is a warning (low salt, low temperature, or the wrong substrate). Bubbling subsides by day 7–10 as the LAB approach the pH where their own activity becomes self-inhibiting.

Foam — a light frothy white — normal in early kombucha (day 1–3) and in vigorous sauerkraut. Disappears as the ferment settles. Heavy foam combined with high temperature is a warning of overactive yeasts.

A new SCOBY layer (kombucha only) — a thin, gelatinous cellulose mat forming on the surface of a kombucha jar at day 3–5. Color may be cream, beige, or brown. Texture is rubbery, smooth. This is correct kombucha activity; the new SCOBY is the bacteria building their own substrate. A patchy or thin new SCOBY is not a safety signal — the kombucha may still be fermenting correctly underneath.

Kahm yeast (flat-white film, no texture) — see §SD-1's kahm exception. Skim, re-submerge solids, move to cooler location. If returning, end the ferment by moving to refrigeration. The flat-white-no-texture description is the operational test: if there is any fuzz, color, or three-dimensional structure to the film, it is not kahm.

Colored mold (green, black, blue, pink, fluffy white-with-structure) — §SD-1 applies. Discard the entire jar. Do not scrape and continue. The visible mold is the fruiting body of mycelium that has been growing for some time.

Surface drying — the brine has evaporated or receded; the top solids are exposed to air. This is the §SD-5 failure waiting to happen. Push solids back under the brine immediately; if the brine level is too low, top up with a 2 % salt solution. If left for more than a day, expect mold or kahm at the next observation.

Tamari pooling on the miso surface — see §3.3. Soybeans liquefying as their proteins hydrolyze. Correct; stir back in or reserve.

Salt cap on the miso surface — should remain firm and white during the first months of the ferment. Cracks or pitting are warning signs (humidity, weight failure).

The cook should look at the surface at every observation date and compare it to the previous observation. A surface that changed in unexpected ways between yesterday and today is data, and the data is usually about temperature, submersion, or the §SD-1 mold question.

§4.4 — Texture

The cook does not need to taste-test every jar daily. Texture is read by looking and (after the ferment is mature enough) by lifting a single piece with clean utensils and pressing it gently.

Crisp — slight resistance to the fingernail or fork — sauerkraut day 10, kimchi day 5, nukazuke at the desired pickle depth. This is the texture target for most vegetable ferments at their decision date.

Soft but structural — gives to pressure without collapsing — sauerkraut day 14, kimchi day 7, nukazuke at longer burial. Still correct; the cook decides whether this is the texture wanted.

Velvety, smooth, almost liquefied — shio-koji day 7, miso month 6. Correct for koji-based ferments. The rice grains in shio-koji have largely broken down; the soybeans in miso have hydrolyzed into paste. This is not a failure; it is the enzymatic ferment's correct end state.

Mushy, collapsed, no structure — sauerkraut day 21+ at room temperature, or vegetable ferment that ran at 28 °C+ for too long. Not unsafe by itself, but past the texture target. Move to refrigeration; use for cooking-in (sauerkraut for soup, kimchi for stew) rather than as a fresh ferment.

Slimy or stringy — vegetable threads pull like cheese — wrong organism (often Leuconostoc mesenteroides gone into rope form, or Bacillus contamination). Operational test: lift a piece with a fork; if a translucent thread follows, that is slime. Discard. §SD-1 frame applies even when no visible mold has appeared. Slime is the bacterial-population signal of trouble.

Grainy (in koji ferments only) — early shio-koji at day 3 is grainy because the koji grains have not yet softened. Continue. Grainy at day 10 means the enzymes never fully activated, probably because the temperature was below 22 °C (§SD-3).

Crumbly or fractured (in miso) — the paste should be cohesive and slightly tacky. Crumbly miso may be over-salted or dehydrated; check the salt cap.

The cook learns texture by handling. The first three or four times the cook lifts a sauerkraut shred, the texture means little. By the tenth jar, the cook can tell with one piece whether the ferment is at day 10 or day 14.

§4.5 — Brine and liquid

The brine — the liquid surrounding the solids — is often the most underread channel in home fermentation. The cook looks at the solids and forgets the brine. But the brine carries most of the cook's information about what is happening chemically.

Submerged solids — §SD-5 is being honored. Continue.

Floating solids — §SD-5 has failed. Push solids back under within minutes if possible; weight them down; if the cook cannot get the solids submerged, the ferment is at risk of surface mold or kahm. This is the moment to intervene, not after the visible failure appears.

Cloudy brine, slightly opaque — correct for an active vegetable lacto-ferment from day 2 onward. The cloudiness is the LAB population suspended in liquid. Clear brine on day 3 of a vegetable ferment is a warning: either the salt was too high (LAB suppressed) or the temperature was too low (chemistry slow).

Evaporated brine, level below the solids — happens when the jar has been vented heavily or stored in dry conditions. Top up with a 2 % salt brine (20 g salt in 1 kg water) to bring the level above the solids. If the brine has evaporated for more than 24 hours, expect §SD-1 issues at the next observation.

Oil layer on the brine surface (vegetable ferment) — should not appear. Vegetable lacto-ferments do not produce oil. If oil is present, something extraneous was added (a recipe variation, or a contaminating ingredient). The cook should be specifically suspicious — this is one of the configurations §SD-2 warns about, because oil produces an anaerobic environment on top of the brine.

Separated brine (vinegar pickle) — sometimes a vinegar pickle's brine separates into layers (vinegar at top, water at bottom) over weeks of storage. Not a safety issue; shake the jar. But also a reminder that vinegar pickles are not fermentation (§3.4) — the separation is a chemistry issue with the dressing, not with microbial activity.

Liquid pooling on the surface of miso — tamari. Correct.

The brine should be checked at every observation. The cook learns to expect clear-then-cloudy in vegetable ferments, sweet-thin-then-thick-opaque in shio-koji, sweet-thin-then-sharp-tart in kombucha. The brine progressions are reliable; deviations from them are signals.

§4.6 — Temperature and speed

Two principles the cook needs to internalize early:

Slow is not failure. A sauerkraut at 12 °C taking 21 days to finish is not a broken sauerkraut. It is a slow sauerkraut. The LAB are still working; they are just working at half-speed. The flavor profile may be slightly cleaner, the texture slightly crisper. The Chapter 1 temperature table makes this explicit (§1.3); the cook should return to it whenever a ferment seems "behind schedule."

Fast is not success. A sauerkraut at 30 °C finishing in 5 days is not a faster sauerkraut. It is a hotter sauerkraut, often softer, sometimes contaminated, frequently with off-flavors that no amount of refrigeration can correct. The decision date arriving early is not the same as the decision date arriving correctly.

The cook chooses speed by choosing the room. A cool pantry corner at 16 °C makes one kind of ferment. A kitchen counter at 22 °C makes another. A warm-summer kitchen at 28 °C makes a third, and the third is at the edge of failure for vegetable ferments while at the optimum for koji (§SD-3 again). The temperature is not a fixed parameter; the cook sets it by placement.

A practical observation: the same ferment in the same kitchen at different times of year is a different ferment. A summer sauerkraut and a winter sauerkraut, made by the same cook with the same recipe, will differ. The cook who notices this is reading temperature correctly.

The instrument: a digital probe thermometer, used once at the start and once mid-ferment. Not constant monitoring. Calibration of intuition, not data logging.

§4.7 — Decision dates

The cook's most important reading is when. When to taste. When to wait. When to refrigerate. When to discard.

When to taste. At the decision date named for each ferment in Chapter 3 — sauerkraut day 10, kimchi day 5, shio-koji day 7, kombucha day 7, nukazuke pickle 4–24 hours depending on vegetable, miso at the 6-, 12-, 18-month marks. Taste with a clean utensil; replace the utensil between tastes. Tasting before the decision date is mostly noise; the ferment is mid-process and the flavor is not yet representative.

When to wait. When the bubbling continues, the brine is correct, the smell is clean-sour, the texture is structural — but the flavor is not yet sharp enough — wait. Continue at room temperature for another 2–4 days, depending on temperature. Tasting daily from the decision date onward is the right cadence. The ferment will tell the cook when it is ready, in the form of a flavor the cook recognizes.

When to refrigerate. When the flavor matches the target. §SD-4. The refrigerator reduces metabolism by roughly a factor of ten; the ferment will continue to develop slowly, but the active fermentation has been arrested. This is the cook's stop button, and the cook should not hesitate to use it once the ferment is at the target flavor. Refrigerated ferments stay safe; over-fermented ferments at room temperature drift.

When to discard. The §SD-1 conditions: visible colored mold, sulfur or putrid smell, slimy texture, salt-below-1.5 % miscalculation (§SD-8), bed unstirred for two weeks (§SD-7), kombucha SCOBY with surface mold. Discard early and discard cleanly. The cook who hesitates to discard a doubtful jar is taking a risk that has no useful upside.

The default for any genuine doubt is discard. The §SD-1 asymmetry-of-cost reasoning applies across every reading channel: the cost of throwing away a jar that might have been fine is small; the cost of keeping a jar that might have been bad is potentially real harm.

§4.8 — Observation log template

The cook learns ferments faster when the cook keeps a log. The log does not need to be elaborate; a small notebook next to the ferments, or a single text file with one line per observation, is enough. The point is the cook's note becomes the source the cook returns to when the next ferment goes a different direction.

Suggested format:

[Date]  [Ferment]  [Day N]  [Room °C]  [Surface]  [Smell]  [Texture if checked]  [Decision]

A worked example over a single sauerkraut:

2026-08-01  Sauerkraut Batch #4   Day 0   22°C  packed/weighted/submerged   clean cabbage/salt   firm   start
2026-08-03  Sauerkraut Batch #4   Day 2   22°C  bubbling vigorous           fresh lactic         (no test)   continue
2026-08-08  Sauerkraut Batch #4   Day 7   22°C  bubbling slowing            clean sour, sharp    crisp     taste tomorrow
2026-08-09  Sauerkraut Batch #4   Day 8   22°C  bubbling minimal            clean sour, balanced crisp     continue
2026-08-11  Sauerkraut Batch #4   Day 10  22°C  no bubbling                 sharp sour           soft-crisp move to fridge — decision date
2026-09-15  Sauerkraut Batch #4   Day 45  4°C   stable in fridge            sharp sour, mellower soft-crisp eating; texture holding

After ten batches in this format, the cook has a personal database. The next sauerkraut is a comparison to the previous nine. The fifth shio-koji is a comparison to the previous four. The cook stops asking "is this normal?" and starts asking "is this different from the last one, and why."

The log is also the cook's protection. If the kombucha that smelled "off" at day 6 turned out to be fine after refrigeration, the log records that — and next time, the cook does not waste a batch on a smell that turned out to be acceptable. If a sauerkraut at day 7 smelled clean and then went sulfur by day 9, the log records that the failure was fast, not slow, and the next sauerkraut is checked more frequently from day 5 onward.

The Notebook recommends the log strongly. It is the one cheap tool that compounds over time.

Closing — what the six channels teach together

The cook who reads the ferment through all six channels — smell, surface, texture, brine, temperature, time — is no longer dependent on the recipe's day-count. They are reading the chemistry directly. They will sometimes call the ferment "done" at day 8; sometimes at day 12; once in a while at day 21. Each decision will be based on what the jar showed, not on what the calendar said.

This is the central skill the Notebook is trying to transmit. Chapter 3 gave the cook six ferments to make. Chapter 4 has given the cook the lens. Chapter 5 — the failure modes catalog — is, in part, a reference for the moments when the lens shows trouble: what the cook saw, what it meant, what the §SD-1 through §SD-8 rules say to do.

If Chapter 4 is internalized, Chapter 5 becomes a sharpened reference rather than a list of fears. The cook stops being afraid of fermentation failure because the failure stops being mysterious. It is observable, before it is irreversible. The cook reads, decides, and acts — and most of the time, the act is the small one (re-submerge, skim kahm, move to fridge). Only the §SD-1 conditions demand the big act, and by the time those appear, the cook has seen the smaller signals leading up to them.

That is what reading the ferment means.

From the Notebook — Chapter 4 closing note

Three reminders the cook will need a hundred times across a fermentation life:

1. Smell first. The nose precedes the eye, the texture, and the calendar. A jar that smells right is almost always right; a jar that smells wrong is almost always wrong.

2. Compare to yesterday. The single observation is data; the day-to-day change is information. A ferment that has not changed since yesterday is either finished or stuck; a ferment that has changed dramatically since yesterday is either accelerating or compromised.

3. Trust the §SD-1 asymmetry. Discarding a fine jar costs a kilogram of cabbage. Eating a compromised jar costs more. When in doubt, the answer is discard.

The rest of the chapter is detail. These three are the spine.

Chapter 4 calibrated the cook's eye, nose, and hand for reading the ferment day by day. Chapter 5 is the reference the cook returns to when that reading shows something off. It is deliberately short on prose and structured for use: each entry names what was observed, what it means, which §SD rule applies, and which of four actions the cook should take. Read once before any ferment; return to it the moment something looks different from yesterday.

§5.1 — How to use this catalog

This catalog is organized by observed signal, not by ferment. The reader who sees mold on a kombucha and the reader who sees mold on a sauerkraut both look up the same entry (§5.2). Inside each entry, ferment-specific notes appear only where the action differs — for example, the §SD-1-limited miso exception is named in §5.2, and koji-specific failure modes have their own §5.12.

The four actions every entry resolves into are:

• Wait — the ferment is on track but appears slow or paused. Continue without intervention.
• Adjust — the cook should change one variable (re-submerge solids, top up brine, move to cooler location) without restarting the ferment.
• Refrigerate — the ferment has reached an acceptable but ambiguous state. Move to cold storage (§SD-4) to arrest development before drift.
• Discard — the ferment cannot be safely consumed. Discard the entire jar, clean per §SD-6, restart with fresh ingredients.

The catalog never proposes an "eat around the bad part" recovery. That class of action is not in this book. The §SD-1 asymmetry of cost applies: discarding a kilogram of cabbage is cheap; eating a compromised ferment is not.

A single-page decision table is at §5.15 for the moment the cook needs only "what do I do right now."

§5.2 — Mold and colored growth

What the cook saw. Green, black, blue, pink, or fluffy white-with-structure growth on the surface of the ferment. The growth has texture, color, or three-dimensional structure — distinct from the flat-white kahm-yeast film described in §5.3.

What it means. Aerobic mold colonization. The visible surface growth is the fruiting body of mycelium that has been growing into the food for some time. Some molds (notably Aspergillus flavus, certain Penicillium) produce mycotoxins that are heat-stable and cumulatively harmful even in small doses.

Rule that applies. §SD-1.

Action: discard. Discard the entire jar. Do not scrape and continue. Wash vessel and tools per §SD-6 before starting a new ferment.

Ferment-specific note — miso (§3.3). Miso is one of the two named structural exceptions to §SD-1 (see Safety Doctrine v1.1, "Named structural exceptions"). The salt cap on top of a correctly-packed miso provides a deliberate physical barrier. If colored mold appears only on the salt cap and the paste below the salt cap smells clean and shows no discoloration, the cook may scrape the affected cap surface 1 cm deep and continue using the paste below. First evaluate structure, color, and smell. Taste only after the named structural exception has clearly passed its visual and smell tests. If the smell or color of the paste below the salt cap is off, or the mold penetrated below the cap, discard the entire tub. This exception is named because it reflects centuries of Japanese practice with the salt-cap structure; it is not an invitation to extend the rule.

Ferment-specific note — kombucha (§3.5). Mold on the kombucha SCOBY or surface means the entire SCOBY and the entire batch are discarded. Do not "rinse the SCOBY." A SCOBY that has hosted mold is contaminated through its structure.

§5.3 — Kahm yeast and surface film

What the cook saw. A flat, white-but-textureless film on the brine surface, no fuzz, no color, no three-dimensional structure. May have a faintly bready or musty smell. Most often appears on vegetable ferments in warm conditions or when submersion (§SD-5) was incomplete.

What it means. A surface yeast colony. Indicates the ferment is at the edge of its safe window but is not by itself a discard signal.

Rule that applies. §SD-1 (kahm-yeast exception).

Action: adjust. Skim the film off with a clean spoon. Re-submerge solids that have floated up. Move the jar to a cooler location. Re-check the surface in 24 hours. If kahm returns, end the ferment by moving to refrigeration (§SD-4) at that point — the ferment is finishing late, and continued waiting will not improve it.

When kahm becomes discard. If the cook is uncertain whether the film is truly flat-white or has any three-dimensional structure, the answer is discard. If kahm yeast and colored mold appear together, the kahm reading was wrong and §5.2 applies. The default for ambiguous surface films is discard.

§5.4 — Slimy texture

What the cook saw. Lifting a piece of vegetable from the ferment with a clean fork, a translucent thread of slime follows; or the brine itself has a stringy, ropy texture. The visible surface may look normal.

What it means. Wrong bacterial population. Often Leuconostoc mesenteroides gone into a rope-forming pathway, or Bacillus species producing extracellular polymer. The ferment is compromised through its biology, not just its surface.

Rule that applies. §SD-1 frame applies even when no visible mold is present.

Action: discard. Slime is a population-level signal. Even if smell and visual surface seem acceptable, the bacterial ecology has gone the wrong direction and will not recover. Discard the jar. Wash vessel per §SD-6. Verify next batch's salt percentage and temperature against §SD-8 and Ch.1 §1.3.

Cause to address before restarting. Most slime failures trace to a single cause: salt below 1.5 % (§SD-8) or temperature held above 28 °C for several days. Re-weigh salt, re-confirm room temperature, before starting a new batch in the same location.

§5.5 — Rotten, sulfur, or garbage smell

What the cook saw. A smell evoking sulfur, rotten meat, garbage, or stale fish. Often noticed within the first 48–72 hours of a vegetable ferment. The cook does not need to be confident in identifying the specific compound; an alarming smell is the data.

What it means. Wrong bacterial population is dominant. Most commonly: insufficient salt allowed Pseudomonas, Enterobacteriaceae, or other spoilage organisms to outpace the LAB before acidification.

Rule that applies. §SD-1 + §SD-8.

Action: discard. Do not attempt to save. The smell is the population's metabolic output; the population cannot be reversed by adding salt or refrigerating. Discard, clean per §SD-6, restart.

Cause to address before restarting. Almost always: salt was below 1.5 %, or temperature was above 28 °C, or submersion (§SD-5) failed. Re-confirm each variable before the next batch. The §SD-8 calculation (salt mass ÷ total mass) should be redone with the scale, not estimated.

§5.6 — Solvent / acetone / nail-polish smell

What the cook saw. A sharp, chemical smell — similar to nail polish remover, paint thinner, or strong solvent. May coexist with otherwise normal-looking surface and brine. Most commonly appears in kombucha that has run too hot, in shio-koji exposed above 35 °C (§SD-3), and occasionally in vegetable ferments where yeast overgrowth has produced significant ethyl acetate.

What it means. Yeast-dominated chemistry producing ethyl acetate, or Bacillus contamination producing related volatile compounds. The ferment has gone past the cook's intended population balance.

Rule that applies. §SD-1 (frame) + §SD-3 for koji ferments.

Action: discard. The solvent smell is not the smell of a finished ferment in any of the six foundations. Even if the rest of the ferment looks acceptable, the chemistry has gone the wrong direction. Discard, clean per §SD-6, restart.

Ferment-specific note — kombucha. A kombucha with solvent smell often correlates with temperatures above 28 °C and/or extended ferment past day 14. Restart at 22–26 °C and respect §SD-4 decision date.

Ferment-specific note — shio-koji. A shio-koji with solvent smell almost certainly experienced temperature above 35 °C (§SD-3). Confirm the temperature setup before the next batch; do not assume the koji culture was at fault.

§5.7 — Too salty or too bland

What the cook saw. Tasting at the decision date, the ferment is either painfully salty (no acid character) or noticeably under-salted (mild and watery, lacking the lactic sharpness expected for the ferment type).

What it means. The §SD-8 calculation was off. Either the salt was over-weighed (too salty) or under-weighed (too bland) at the start.

Rule that applies. §SD-8.

Action depends on which direction:

• Too salty + ferment otherwise correct (clean smell, no mold, no slime): The ferment is safe. The LAB simply ran more slowly than at the intended salt percentage. Move to refrigeration; use the ferment as a strongly-salted condiment rather than as a daily eating pickle. Re-weigh salt for the next batch.

• Too bland + salt was actually above 1.5 %: The ferment is safe but under-developed. Continue at room temperature for another 5–7 days if other signals are normal. If still bland after that, the salt was high enough to slow the LAB but not high enough to flavor the result; the cook chose a difficult ratio. Move to refrigeration and use for cooking-in. Re-weigh salt for the next batch.

• Too bland + salt was actually below 1.5 %: §SD-8 says discard. The ferment is in the danger zone regardless of how it tastes. Discard, clean per §SD-6, restart with correctly-weighed salt.

Tools to address before restarting. A digital kitchen scale accurate to 1 g. Volumetric salt measurements (teaspoons) are a known cause of this failure mode (see Ch.1 §1.2 source essay).

§5.8 — No bubbles, no progress, ferment looks dormant

What the cook saw. Day 3 of a vegetable ferment: no bubbling visible against the glass, no rising brine, no change in smell from day zero. The jar looks like the cook just packed it.

What it means. One of three things: the temperature is below the active LAB window; the salt is above 5 %; or the LAB inoculation was too low (the cabbage was washed too thoroughly, or was stored in a sterile environment before the ferment).

Rule that applies. Ch.1 §1.3 temperature window, §SD-8 salt range.

Action: adjust, then wait.

1. Check the room temperature. If below 16 °C, move the jar to a slightly warmer location (18–22 °C) and check again in 48 hours.
2. Re-confirm the salt percentage by mental arithmetic: salt mass ÷ total mass. If above 5 %, the LAB are suppressed and the ferment will run slowly but will eventually finish.
3. If both temperature and salt were correct, the ferment may simply have a slow start. Wait another 5–7 days at the corrected location. If still no bubbling by day 10, the ferment did not establish; discard and restart with a small amount of brine from a successful previous batch as a starter culture.

When dormancy becomes discard. If at day 14 with corrected temperature and salt there is still no bubbling and no acid character, the ferment did not establish. Discard. Do not attempt to "kickstart" by adding sugar or warm water at this stage; the cook does not control which organism wakes up.

§5.9 — Ferment ran too fast, now over-acidic and mushy

What the cook saw. Day 5 (rather than day 10): the brine is fully cloudy, bubbling has subsided, the vegetable is soft and breaking down, and the smell is sharply sour, approaching vinegar. The cook had not yet planned to refrigerate.

What it means. The ferment ran at a higher temperature than intended, and the chemistry has moved past the decision date the cook had on the calendar. The ferment is not unsafe — it is finished, just earlier and harder than planned.

Rule that applies. Ch.1 §1.3 (temperature window upper bound) + §SD-4 (decision date as cook's choice, not ferment's).

Action: refrigerate. Move the jar to cold storage immediately. The ferment will arrest; the texture and flavor are locked at day 5. Use the result for cooking-in rather than as a fresh pickle (sauerkraut into soup, kimchi into stew). Re-confirm temperature setup before next batch.

When fast becomes discard. If alongside the fast ferment the cook also detects off-smells (§5.5), slime (§5.4), or surface mold (§5.2), then the fast ferment was also compromised. §SD-1 applies; discard.

§5.10 — Brine loss, exposed solids

What the cook saw. Brine level has fallen, and solids are now floating above the brine surface or exposed to air. The cook may notice this at day 3 (jar vented too aggressively) or day 7 (slow evaporation).

What it means. §SD-5 is breaking. Surface mold or kahm yeast (§5.2, §5.3) is becoming likely within 24–48 hours if the cook does not intervene.

Rule that applies. §SD-5.

Action: adjust, immediately.

1. If solids are floating: push them back under the brine. Use a clean fork or weighted glass.
2. If brine level is too low to cover solids after re-submerging: top up with a fresh 2 % brine solution (20 g salt dissolved in 1 kg filtered water). Pour gently around the solids until the level rises above them.
3. Re-weight the solids to keep them submerged. The cook may need a heavier weight or a different vessel.
4. Re-check in 24 hours. If brine continues to fall, the jar's seal is over-venting and may need replacement.

When brine loss becomes discard. If solids have been exposed for more than 24 hours and surface mold has appeared, §SD-1 applies; discard. The recovery is preventive — re-submerge before the failure shows; once §SD-1 conditions are visible, the §5.2 entry applies.

§5.11 — Alcohol or vinegar character in a vegetable lacto-ferment

What the cook saw. Tasting a sauerkraut or kimchi at the decision date, the cook notes an alcoholic note (like very low-alcohol beer) or a sharp vinegar character (much sharper than the expected lactic sour).

What it means. Either yeast overgrowth (producing ethanol) or acetic-acid bacteria contamination (producing vinegar). Both indicate the LAB did not dominate the chemistry as intended.

Rule that applies. Ch.1 §1.5 (interaction of variables); §SD-1 frame; §SD-2 relevant when the acidification pathway itself has faltered (a ferment that has not driven pH below 4.6 by day 3–5 is not just yeasty; it is potentially in the anaerobic + low-acid zone the botulism rule guards against).

Action depends on severity:

• Mild alcohol note, otherwise clean ferment: The ferment is safe but has gone slightly yeasty. Move to refrigeration to arrest further development. Use as-is.
• Strong alcohol note, sharp vinegar character, or any off-smell alongside: Discard. The chemistry has gone in a direction outside the lacto-ferment frame; the ferment is no longer the food the cook intended.

Cause to address before restarting. Yeast overgrowth is usually a temperature issue (above 26 °C) plus a too-loose seal. Acetic-acid contamination is usually a sealing issue plus extended ferment time. Tighten temperature control and check decision date before next batch.

Boundary note. A vegetable lacto-ferment should not taste like vinegar or like kombucha. If the cook is making a vegetable pickle and the result tastes alcohol-adjacent, the ferment crossed into territory that does not belong to its category. The Notebook discusses kombucha (intentionally alcohol-and-vinegar adjacent) in §3.5; a sauerkraut should not arrive there.

§5.12 — Koji-specific failures (shio-koji, miso)

Koji ferments have failure modes that the vegetable-ferment cook will not have encountered. The §SD-3 temperature window (22–35 °C, distinct from the vegetable 18–22 °C window) is the root of most.

Shio-koji failed to thicken / liquid stayed clear at day 10. Temperature was below 22 °C; the koji enzymes did not activate sufficiently. Adjust: move to a yogurt maker or sous-vide bath at 28 °C for an additional 7 days. If still clear after 7 more days, the koji culture was likely weak or expired; discard and restart with fresh koji.

Shio-koji turned pink, red, or grey-green. Bacillus or yeast contamination from temperatures above 35 °C. §SD-3. Discard. The discoloration is not flavor character; it is the wrong organism.

Shio-koji surface developed mold. §SD-1 + §SD-5 (the jar was probably not loosely sealed). Discard.

Miso developed strong ammonia smell. Long ageing in too-warm conditions (above 30 °C for extended periods), or Bacillus contamination at start. Discard the tub. This is one of the most serious miso failures because ammonia indicates protein breakdown in the wrong direction.

Miso developed pink or red discoloration in the paste below the salt cap. §SD-1. Discard the tub. Unlike surface mold (§5.2 miso exception), discoloration of the paste itself is a population-level signal.

Miso tamari pooling on top is correct. This is the soy proteins liquefying. Not a failure. Stir back in or reserve.

§5.13 — Kombucha-specific failures

Kombucha's mixed-culture chemistry (yeast + Acetobacter) creates failure modes the lacto-ferment cook will not have seen.

Mold on the SCOBY or jar surface. §SD-1. Discard SCOBY and entire batch. Do not "rinse" the SCOBY. The mold has likely penetrated the SCOBY's cellulose structure.

Strong solvent or putrid smell. §SD-1 + §5.6 (solvent entry). Discard. Likely temperature above 28 °C or wrong organism dominant from start.

Slimy, ropy texture in the liquid. Mother-of-vinegar-style overgrowth or yeast bloom. Discard the batch. The cook may keep the SCOBY if it has not been in direct contact with the slime, but a safer default is to start with a fresh SCOBY.

Vinegar character is overwhelming and the kombucha is hyper-sour. The ferment ran past day 14 at warm temperature. Not unsafe; use as a vinegar substitute. Compost the SCOBY if it has aged out and start fresh.

SCOBY is dark, leathery, or strangely shaped. A SCOBY can look strange to the new fermenter and still be functional. The §SD-1 question is mold and color (green, black, blue, pink). A brown, beige, or cream SCOBY with no fuzz is normal. Continue.

New SCOBY layer is thin, patchy, or absent at day 7. Not by itself a safety signal. The primary ferment may still work. Check pH if possible (should be below 4.0 by day 5). If pH is correct and smell is clean kombucha, continue. If pH is not dropping and smell is not characteristic of kombucha, discard.

§5.14 — Nukazuke-specific failures

The nukazuke bed is the §SD-7 ferment, and its failure modes are different again from single-batch ferments.

Surface mold (any color). §SD-1 + §SD-7. Nukazuke is the second of the two named structural exceptions to §SD-1 (see Safety Doctrine v1.1) — because the bed is a continuous-process ferment rather than a single batch. Limited recovery available. Scoop off the affected surface 2–3 cm. First evaluate structure, color, and smell of the bed below. Taste only after the visual and smell tests have clearly passed. If clean and unaltered, the bed continues. If smell is off or mold penetrated deeper, discard the entire bed. The §SD-7 cost of discarding a long-tended bed is real; the cost of eating a compromised bed is higher. As with the miso exception, this is named because it reflects centuries of nuka-bed practice, not a softening of the discard-first default for ferments without the same continuous-process structure.

Strong yeast bloom (powdery white surface). Bed not stirred for several days; temperature too high. Adjust: stir thoroughly, reduce temperature (move to cooler corner), add a small replenishment of fresh nuka. If bloom returns within a week, the bed may need partial replacement (20–30 % fresh nuka).

Off-smell (rotten, fishy, putrid). The bed has been compromised, likely by a vegetable that introduced wrong organisms, or by extended unstirred time. Discard the bed. Start fresh. This is the §SD-7 worst case and the reason daily stirring matters.

Bed has become too sour (excessive lactic acid). Over-fermented; not enough fresh nuka added; vegetable burials too frequent. Adjust: add fresh nuka + salt (about 10 % of bed mass); reduce vegetable burial frequency for a week.

Bed has become too watery. Repeated vegetable burials added water without enough nuka replenishment. Adjust: add fresh dry nuka to absorb. Stir thoroughly.

Cook has been away 2 weeks and the bed was unstirred. §SD-7. The bed is likely compromised. Assess smell and surface carefully. If any §SD-1 conditions are visible, discard. If smell is clean and no surface bloom or mold, the bed may have survived in refrigeration if the cook had moved it before leaving; if it was at room temperature for the duration, default to discard and restart with fresh nuka.

§5.15 — Decision table: wait / adjust / refrigerate / discard

A single-page reference for the moment the cook needs only "what do I do right now."

From the Notebook — Chapter 5 closing note

Two things to remember when this catalog is open:

1. The decision is action, not diagnosis. The cook does not need to identify the specific organism that produced the smell or the surface growth. The cook needs to choose among the four actions. Wait / adjust / refrigerate / discard.

2. The default for ambiguity is discard. Every entry above can be triangulated to one of the four actions, but the catalog cannot anticipate every observation the cook will encounter. When the observation does not match an entry, and the cook is genuinely uncertain, the answer is discard. This is the §SD-1 asymmetry of cost expressed as a routine.

The catalog is for re-reading at the moment of decision, not for memorization. A small notebook open next to the ferments, bookmarked at §5.15, is the cook's friend.

Chapter 5 was the dictionary the cook returns to when something looks wrong. Chapter 6 is the rhythm the cook lives with when things are going right. Some ferments — the nukazuke bed, the kombucha mother, the long-aged miso tub — are not one-time projects. They are relationships, and they ask for attention at scales the rest of the book has only hinted at. This chapter sets the cadence. Daily for the nuka bed. Weekly for the kombucha. Seasonal for the miso. And, when the cook eventually has to walk away — for travel, for changed circumstances, for the decision that a ferment has run its course — Chapter 6 also names how to do that honestly. Including the chapter's central ethical claim, which the book has been moving toward since the Introduction: discard is not failure; it is maintenance.

§6.1 — A living ferment is not a pet, but it asks for attention

A nukazuke bed in the corner of the kitchen will not greet you when you come home. It does not know your name. It is not happy when you stir it and it is not sad when you go away. The cook who romanticizes the ferment as a companion is making a category error that will eventually distort their decisions about it — either by hesitating to discard a bed that should be discarded, or by attributing flavor changes to the bed's "mood" rather than to the temperature or salt or burial schedule.

The ferment is a managed ecosystem. The cook's role is closer to a gardener than to a pet-owner. A gardener watches a garden, intervenes at known points (water, prune, weed, harvest), and accepts that the garden has its own logic separate from the gardener's affection. The garden does not need to be loved. It needs to be tended.

This is also what makes the relationship sustainable. A cook who feels emotionally attached to a kombucha SCOBY is the cook who will try to "save" a moldy SCOBY past the §SD-1 discard signal — and will be hurt. A cook who treats the SCOBY as a chemistry tool that has produced excellent results until today, but is now compromised, can discard cleanly and start a fresh one without grief. The work continues; the cook's stewardship continues; the specific SCOBY does not. Nothing about that is a tragedy.

But attention is real. A nukazuke bed left unstirred for two weeks (§SD-7) is no longer the bed it was. A kombucha jar left at 30 °C through a heat wave is no longer the kombucha the cook intended. A miso tub checked once in a year is missing the seasonal information that would have told the cook to move it before summer. The ferments do not punish neglect. They simply drift, and at some point in the drift they cross a §SD-1 line, and the cook has to discard a relationship that could have been preserved with a small consistent intervention along the way.

The frame this chapter holds: attention without sentimentality. The cook brings rhythm and observation to the ferment. The ferment provides the cook with foods that fresh ingredients cannot make. The exchange is honest. Neither party owes the other anything beyond what the chemistry requires.

§6.2 — Daily relationships

Three ferments in the book ask for daily attention. The cook should treat them as routine — paired with another daily habit (morning coffee, evening dish-washing) so the attention does not depend on the cook remembering separately.

Nukazuke bed. Once daily, stir the bed thoroughly by hand or with a wooden paddle. Bring surface material into the interior. Take fifteen seconds to smell: the bed should smell of fermented rice bran with a faintly earthy edge — clean, slightly sour, never putrid. Check the surface for any film or discoloration. Replace any vegetables that have been buried long enough to reach the cook's desired pickle depth (cucumber 4–8 hours, daikon 12–24, cabbage 24–48). The whole interaction is about three minutes once it becomes routine.

Shio-koji during the active 7-day window. Once daily, stir with a clean spoon to redistribute the enzymes and prevent surface drying. Brief smell check: shio-koji should smell faintly sweet and slightly sour, never solvent-like (§5.6). Check that the koji is staying within its 22–35 °C temperature window (§SD-3). Twenty seconds.

Any active vegetable brine ferment in the first 7 days. Walk past the jar, look at the brine level (§SD-5 submersion), confirm bubbling is present (§4.3), and smell briefly through the lid. Ten seconds. The point is not detailed reading; the point is catching changes before they become problems.

The cook who builds daily attention into a routine never has to "remember to check on the ferments." The check happens automatically, attached to a more reliable habit. The total daily fermentation time, with a mature nuka bed and an active brine ferment in progress, is rarely more than five minutes.

§6.3 — Weekly relationships

Some ferments operate on a weekly cadence. The cook does not need to look at them daily; they need to look at them on a stable weekly schedule.

Kombucha primary ferment. A kombucha jar is checked at the day-7 decision date. Before that, the cook may glance at it but does not need to taste or intervene. The weekly check determines whether to bottle (move to secondary, refrigerate) or extend (continue at room temperature for sharper character). If the cook brews kombucha continuously, the weekly cycle becomes the rhythm: taste, bottle, restart, repeat.

Refrigerated ferments. Sauerkraut, kimchi, and other completed ferments now in cold storage benefit from a weekly check. The cook tastes a small amount, confirms the texture has not collapsed (§4.4), and confirms the smell remains clean-sour without drifting toward off-character. A sauerkraut in the refrigerator for two months is normally still excellent; a sauerkraut for six months may be developing flavors the cook should know about before service.

Brine top-ups. Long-fermenting jars (nukazuke not in scope here — that is daily) may lose brine slowly. A weekly look at brine levels for any ferment older than three weeks at room temperature catches §SD-5 failures before they become surface mold.

Decisions to refrigerate something that has finished. A cook who has been waiting for a sauerkraut to reach a specific sourness may, on the weekly check, decide the day has arrived. Move to refrigeration immediately. The §SD-4 stop button does not need to wait for an arbitrary day-count to be reached; the weekly check is the decision opportunity.

The weekly check should also include any kombucha SCOBY hotel (a jar where extra SCOBYs are kept in mature vinegar between brewing cycles) — refresh the brine if it has gone too dry, discard any SCOBY that has developed §SD-1 conditions.

§6.4 — Long relationships

Some ferments work on a scale where daily or even weekly attention is wrong — what they need is the discipline of leaving them alone.

Miso, 6–24 months. The cook checks monthly, not daily. Monthly is enough to catch surface yeast or kahm before it establishes, to confirm the salt cap is firm, and to confirm no §SD-1 conditions (Doctrine v1.1 miso exception still requires the operational test). A cook who opens the miso weekly is introducing oxygen and disrupting the surface seal. A cook who never opens it for a year may miss a surface bloom that would have been easy to skim three months earlier. Monthly is the calibrated cadence.

Long-aged nukazuke bed (year two onward). Once a mature bed has been established and the daily-stir routine is in place, seasonal checks become more important than daily ones. The cook should refresh 10–20 % of the bed twice a year (typically at the start of summer and the start of winter, when the kitchen temperature shifts). The annual cycle keeps the bed from accumulating excessive salt drift or texture fatigue. This is in addition to the daily stir, not a replacement.

Sourdough cultures (out of scope for this Notebook but adjacent). Mentioned here because cooks who have a sourdough may also have other ferments going. A sourdough that is fed weekly (refrigerated) is on a different cadence from one being fed daily at room temperature. The principles transfer: respect the cadence the system asks for; do not over-intervene; do not under-attend.

The discipline of NOT looking. This is the harder skill. A cook who opens a miso tub every week, "just to check," is harming the ferment more than helping it. The salt cap loses integrity. Oxygen enters. The §SD-1 risk goes up, not down. The same is true of long-aged kombucha vinegar, of nuka beds that have settled into a stable population. Some attention is care; too much attention is interference. The cook learns the difference by feel, over the years the ferments have been part of the kitchen.

§6.5 — Absence: travel, illness, change of circumstance

The cook will, at some point, need to leave the kitchen. A week of travel. An illness. A move. A change of seasons that makes the kitchen unsuitable for a ferment that was happy there in spring.

The nukazuke bed. Up to 7 days of refrigeration (covered, in the fridge, at 4 °C) keeps a mature bed alive without active attention. Beyond 7 days, the bed stresses. Beyond 14 days, the bed is likely compromised (§SD-7). For longer absences: give the bed to another cook who knows the routine, or accept that it will need to be restarted. The choice is honest; the cook who has been tending a 3-year bed will feel the loss of restarting, and the loss is real. It is also not a failure of stewardship.

Kombucha. A SCOBY in mature vinegar can rest for weeks without active brewing. Bottle the in-progress batch, refrigerate it; place the SCOBY in a clean jar with enough mature kombucha to cover it; refrigerate or leave at cool room temperature. The SCOBY survives. The cook resumes brewing when the cook returns.

Refrigerated ferments. Already in cold storage; absence does not affect them.

Active vegetable lacto-ferments mid-process. If the ferment is at day 3 and the cook is leaving for two weeks, the safest move is to refrigerate immediately — accept that the result will be milder than intended, but it will be safe. The §SD-4 stop button is also the cook's "I have to leave" button. The ferment will continue to develop slowly in cold storage; the cook returns to a finished ferment, not a lost one.

Active koji ferment. If a shio-koji is at day 3 and the cook must leave for a week, refrigeration arrests it; the enzymes will not have fully done their work, but the result is usable. A shio-koji that has been left at active temperature for two weeks unattended will not be a good shio-koji.

Forgetting. This is the version of absence the cook does not plan for. A nuka bed not stirred for two weeks because life got in the way. A kombucha that ran for three weeks because the cook lost track of which jar was which. The §SD-1, §SD-7, and §5.14 rules still apply. The cook does not pretend the time did not pass. The cook checks honestly. If the bed is compromised, it is discarded. The next bed will be easier to start because the cook now knows the rhythm.

§6.6 — Restarting without guilt: discard as maintenance, not failure

This is the chapter's central ethical claim, and the Notebook's. Read it carefully.

A cook who has been tending a nukazuke bed for two years, who discovers on a Tuesday morning that the bed has gone unmistakably wrong — pungent off-smell, surface mold penetrating deeper than the §5.14 exception allows, a paste below the surface that no longer smells like the cook's bed — has not failed. The cook has maintained the bed up to the point where the bed could no longer be maintained, and at the moment of decision the cook acted correctly. The bed is discarded. The kitchen is cleaned. A new bed is started.

This is not the failure mode the cook needs to fear. The failure mode the cook needs to fear is the opposite — the cook who could not bring themselves to discard, who scraped a little deeper than the §SD-1 exception allowed, who told themselves the smell "wasn't that bad," who ate from a bed that should have been thrown out. That cook is the one who is eventually hurt.

The asymmetry of cost the book has stated many times applies here as a guiding principle for the cook's relationship with the ferment itself, not just for individual jars. A cook who can discard cleanly is a cook in control. The relationship with the bed is not the bed; it is the cook's practice of tending. The bed is replaceable. The practice is not.

This is also what makes the cook willing to keep ferments going. If discard were felt as failure, the cook would unconsciously avoid starting ferments at all, or would over-extend ferments past the moment they should be stopped, just to avoid the loss. Both behaviors are bad for the ferments and bad for the cook's safety. The reframe — discard is maintenance — is what lets the cook tend ferments without ego involvement. The first nuka bed will be discarded eventually. The fifth one will too. The cook who has internalized this can run continuous ferments for decades.

When the cook does discard, the action should be unceremonial. Empty the jar into the compost (or trash, if §SD-1 conditions argue against compost — moldy ferments are not necessarily safe for the compost ecosystem either). Wash the vessel in hot soapy water (§SD-6). Replace any wooden tool that touched colored mold. Wait 24 hours before starting a new batch in the same location. Move on.

The cook keeps a log (§4.8). The log records discards alongside successes. Over the years, the log becomes the cook's professional memory: this bed went two years; this one eight months; this one only a week, because the kitchen was too hot that summer. The log is not a scorecard of failures and successes. It is the cook's working knowledge of the medium.

§6.7 — Passing a culture forward: when sharing is safe, when not to share

Some of the foundation ferments produce surplus culture that can be passed to other cooks. A kombucha SCOBY produces a new daughter SCOBY with every batch. A nukazuke bed can be partitioned — the cook gives a third of the bed to a friend who is starting their own. A miso tub at year three can yield a quantity of starter that helps a friend's first miso get underway.

When sharing is safe. The cook may share a culture when:

• The source ferment has no §SD-1 conditions and has not had any in the recent past.
• The receiving cook understands the maintenance rhythm. (A nuka bed handed to a cook who will not stir it daily is a bed being given a death sentence; the gift is not a kindness.)
• The receiving cook understands the discard rules. The book's §SD-1 through §SD-8 apply to inherited cultures the same way they apply to home-started ones.
• The transfer happens cleanly. Sterile jar, clean transfer, clear handoff instructions.

When NOT to share. The cook should not share:

• A culture that has recently had §SD-1 conditions, even after surface treatment. The receiving cook inherits an unknown risk.
• A nuka bed during a §SD-7 violation window (cook traveling, bed un-stirred). Even if the bed seems fine, the next 48 hours will tell the story; do not pass the uncertainty to another cook.
• A culture to a cook who has not been told the §SD-1 discard rules. Sharing a SCOBY without sharing the rules for when to discard it is incomplete; the receiving cook may eventually be hurt by the gift.
• Any culture out of curiosity or sentimentality, without confirming the receiving cook actually wants the responsibility.

A culture handed forward is also a transfer of practice. The cook giving the gift is, in effect, asking the receiving cook to maintain a rhythm of attention. This is a non-trivial ask. The Notebook recommends a hand-written one-page note accompanying any handed-forward culture: the salt percentage, the temperature window, the daily/weekly cadence, the §SD-1 conditions to watch for. The receiving cook can then decide whether to accept.

§6.8 — Closing: fermentation as managed time

The Notebook has, across six chapters, taught the cook to read salt percentages, hold temperature windows, watch decision dates, observe surfaces and smells and textures, look up failure modes, and tend living systems on cadences from daily to seasonal. All of these skills, when collected, are aspects of the same craft: managed time.

A 10-day sauerkraut is the cook managing 10 days. A 6-month miso is the cook managing 6 months. A 5-year nuka bed is the cook managing 5 years of small daily decisions. None of the recipes in this book — and none of the recipes at terumimorita.com, and none of the recipes the cook will encounter in the rest of fermentation literature — are really about ingredients. They are about timescale. The cook chooses the timescale; the chemistry decides what happens at that timescale; the cook reads, decides, and acts.

This is also what makes fermentation different from the rest of cooking, and what is rewarding about it. A soup is finished in twenty minutes. A roast is finished in two hours. A fermented preparation has not yet started after two hours, and may not finish until the cook's child is grown. The cook who has internalized the Atlas chapter framing (Ch.11 of the free Atlas) understands the historical scale on which this practice operates — ten thousand years of human cooks managing time using salt, acid, dehydration, and microbes.

The Notebook is one practical tool in that ten-thousand-year tradition. It teaches the rules that make a small kitchen-counter version of the practice safe and repeatable. It does not claim originality; the rules are old, and have been written about in many other books in many other languages. What it claims is honesty. The salt percentage is given. The temperature window is given. The discard rule is given. The exceptions are named. The cook who follows the book and tends the ferments will, over the years, become competent at a practice that is older than agriculture and that no algorithm can compress into a shortcut.

There is no shortcut. The cook chooses the timescale and waits. That is the entire teaching.

From the Notebook — Chapter 6 closing note

The book began with a claim in the Introduction: fermentation is not a flavor pursuit; it is environment design. It ends with the same claim from the opposite direction: fermentation is the cook's relationship to time at a scale that the rest of cooking does not require. Both claims are the same observation in different language. The cook designs the environment; the environment unfolds across time; the cook tends what unfolds.

Three final reminders the cook will carry forward:

1. Attention without sentimentality. The ferment is a managed ecosystem. The cook brings rhythm and observation. The chemistry does the work.

2. Discard is maintenance, not failure. A cook who can discard cleanly is a cook in control. The asymmetry of cost (§SD-1) applies to every relationship the cook tends.

3. The timescale is the cook's choice; the chemistry is the chemistry. A 10-day ferment and a 5-year ferment are the same craft at different scales. The cook chooses the scale, then respects what the chemistry does at it.

The book ends here. The ferments continue.

The Appendix is a working reference. It exists to be flipped to in the moment the cook needs a number, a definition, or a source. It is not meant to be read cover-to-cover. The tables here duplicate (in compact form) the specifications introduced across Chapters 1–6, so the cook does not need to flip through the body chapters to find a salt percentage or a temperature window. Sections A.4 (observation log) and A.5 (decision table) are formatted as printables — designed to be printed on a single A4 page and clipped near the ferments. Section A.7 names the external safety sources the book leans on, so the cook can verify the rules in this Notebook against authoritative public-health and home-food-preservation literature.

A.1 — Tools and instruments

The kitchen does not need much equipment for the six foundation ferments, but the equipment it does need should be reliable.

Kitchen scale, accurate to 1 g. Non-negotiable. The salt percentage is the single most important variable in vegetable fermentation (§SD-8), and the difference between a safe and a dangerous salt level is sometimes ten grams. Volumetric measurements (teaspoons, cups) are unreliable for salt because the bulk density varies between brands by a factor of two (Ch.1 §1.2 documents the Diamond Crystal / Morton / table-salt example). A digital kitchen scale that reads to 1 g costs roughly the same as a single bottle of olive oil and is in use for every ferment for the rest of the cook's life. Buy first.

Glass jars, wide-mouth, 1 L for vegetable ferments. Wide-mouth glass is the standard. Glass is non-reactive, lets the cook see the brine and surface, and washes clean between ferments. Plastic jars work but stain and absorb odors over time. Ceramic crocks (the traditional German Gärtopf) are excellent for larger sauerkraut batches but are an upgrade, not a starting necessity.

Fermentation weight. A glass disc, a smaller jar filled with water, or a cabbage leaf folded over the solids with a weight on top — anything that keeps solids submerged below the brine (§SD-5). Dedicated fermentation weights are widely sold; a clean smooth stone works too.

Digital probe thermometer. Useful at the start of every ferment to confirm the temperature window (§4.6 + Ch.1 §1.3). Especially important for koji ferments where the 22–35 °C window (§SD-3) is narrow. A simple kitchen-counter probe thermometer is sufficient; the cook does not need a continuous monitor.

pH strips or a pH meter (optional but useful). Vegetable ferments correctly acidify below pH 4.0; the §SD-2 botulism boundary is pH 4.6. The cook can confirm acidification with pH strips that read in the 3.0–5.0 range, available at brewing supply shops. A pH meter is more precise; not necessary unless the cook is troubleshooting a difficult ferment.

Labels and date tape. Painter's tape and a marker. The cook writes the ferment name, the start date, and the calculated salt percentage on the tape. This is the input to the observation log (§A.4) and the cook's protection against confusion when multiple ferments are running.

For koji ferments specifically (§3.2, §3.3). A way to hold the jar at 25–28 °C — a yogurt maker, a sous-vide bath set to 28 °C, or a turned-off oven with the light on. This is the most demanding piece of kit in the book and is only necessary for shio-koji and miso starts.

For nukazuke specifically (§3.6). A ceramic or wooden tub of at least 3 L capacity, with a tight-fitting lid. Dedicated to the bed; not for use in other cooking.

A.2 — Ratio quick table

All ratios are by weight. Salt percentages are calculated as salt mass ÷ total mass × 100. Volumetric measurements are not used in this book (Ch.1 §1.2).

Ferment                 Salt %       Substrate                          Water         Notes
─────────────────────────────────────────────────────────────────────────────────────────────
Sauerkraut              2.0%         1 kg shredded cabbage              none          §3.1
Kimchi-style            2.0–2.5%     1 kg cabbage + aromatics           none          variation
Nukazuke bed            7–10%        1 kg rice bran (toasted)           1 kg          §3.6
Shio-koji              11.1% total  200 g rice koji                    280 g         §3.2
                                     + 60 g salt
Miso-style (small)     12% total    500 g cooked beans                 ~150 g        §3.3
                                     + 500 g rice koji
                                     + 150 g salt
Vinegar pickle         not ferm.    500 g vegetable                    250 g + 250 g §3.4
                                     + 10 g salt                        vinegar+water boundary
Kombucha               0% salt      5–8 g tea                          1 kg water    §3.5
                                     + 70 g sugar
                                     + 100 g starter liquid             SCOBY

The cook re-runs the percentage calculation before every batch. The kitchen scale is the only reliable instrument; the recipe number is a target, not an estimate.

A.3 — Temperature quick table

Temperatures in this book are the windows the author finds manageable in a home kitchen and that produce reliable results across seasons. They are not the published absolute bounds; many ferments work safely outside these ranges. See §A.7 for the framing rationale (the cook can ferment colder for safer/slower results; warmer is a known risk territory for vegetable ferments).

Ferment / state          Working °C    Lower OK            Upper risk
──────────────────────────────────────────────────────────────────────
Vegetable lacto-ferment  18–22 °C      12 °C (slow)        > 28 °C
Kimchi initial           18–22 °C      15 °C (slow)        > 25 °C (mush)
Kimchi cold storage      0–4 °C        colder OK           n/a
Nukazuke bed (active)    15–22 °C      cellar OK           > 28 °C
Shio-koji activation     25–28 °C      22 °C (slow)        > 35 °C (§SD-3)
Shio-koji storage        0–4 °C        colder OK           n/a
Miso (long ferment)      15–25 °C      cooler OK (slower)  > 30 °C
Kombucha                 22–26 °C      18 °C (slow)        > 28 °C
Vinegar pickle           0–4 °C        always cold         > 4 °C unsafe
Refrigerator storage     0–4 °C        target              > 4 °C drift

Reminder (§SD-3): koji and vegetable lacto-ferments have different temperature windows. A 28 °C sauerkraut is at the failure edge; a 28 °C shio-koji is at optimum. Hold each ferment to its own window.

A.4 — Observation log printable

A single-page printable for clipping near the ferments. The cook fills one line per observation.

┌──────────────────────────────────────────────────────────────────────────────────┐
│  FERMENT OBSERVATION LOG                                  Batch ID: _____________│
├──────────────────────────────────────────────────────────────────────────────────┤
│  Ferment name: ______________________   Start date: ______ Salt %: __________   │
│  Ferment type:  ☐ sauerkraut  ☐ kimchi-style  ☐ shio-koji  ☐ miso              │
│                 ☐ vinegar pickle (not ferm.)  ☐ kombucha   ☐ nukazuke           │
│  Target decision date: ______                                                    │
├──────┬─────┬──────────┬───────────────┬───────────────┬───────────────┬────────┤
│ Date │ Day │ Room °C  │ Surface       │ Smell         │ Texture/taste │ Action │
├──────┼─────┼──────────┼───────────────┼───────────────┼───────────────┼────────┤
│      │     │          │               │               │               │        │
├──────┼─────┼──────────┼───────────────┼───────────────┼───────────────┼────────┤
│      │     │          │               │               │               │        │
├──────┼─────┼──────────┼───────────────┼───────────────┼───────────────┼────────┤
│      │     │          │               │               │               │        │
├──────┼─────┼──────────┼───────────────┼───────────────┼───────────────┼────────┤
│      │     │          │               │               │               │        │
├──────┼─────┼──────────┼───────────────┼───────────────┼───────────────┼────────┤
│      │     │          │               │               │               │        │
├──────┴─────┴──────────┴───────────────┴───────────────┴───────────────┴────────┤
│  Decision date taken: ______  Final action: ☐ refrigerated  ☐ discarded         │
│  Result tasted:                                                                  │
│  Note for next batch:                                                            │
└──────────────────────────────────────────────────────────────────────────────────┘

Action codes: W = wait · A = adjust · R = refrigerate · D = discard. The codes match the four-action vocabulary of Chapter 5.

Print notes: the log is sized for A4 portrait at monospace 10–11 pt. For a larger writable field, print at A4 landscape and increase the row height. Keep one log per batch; do not reuse across batches.

A.5 — Decision table printable

A compact one-page version of Ch.5 §5.15. Designed to be printed and clipped next to the observation log.

┌──────────────────────────────────────────────────────────────────────────┐
│  DECISION TABLE — wait / adjust / refrigerate / discard                  │
├──────────────────────────────────────────────────────────────────────────┤
│  SIGNAL                                  RULE          ACTION            │
│  Bubbles vigorous, smell clean, day 3    (correct)     WAIT              │
│  Bubbles slowing, sour, crisp, day 10    §SD-4         REFRIGERATE       │
│  No bubbles day 3                        §1.3 temp     ADJUST → wait     │
│  No bubbles day 14                       §SD-8         DISCARD           │
│  Flat-white film, no fuzz                §SD-1 kahm    ADJUST (skim)     │
│  Colored mold (green/black/blue/pink)    §SD-1         DISCARD           │
│  Slimy / stringy texture                 §SD-1 frame   DISCARD           │
│  Sulfur / rotten / putrid smell          §SD-1 + §SD-8 DISCARD           │
│  Solvent / acetone / nail-polish smell   §SD-1 + §SD-3 DISCARD           │
│  Mild alcohol note, clean otherwise      (yeast)       REFRIGERATE       │
│  Strong alcohol / vinegar character      §SD-1 frame   DISCARD           │
│  Too salty, ferment otherwise correct    (over-salt)   REFRIGERATE       │
│  Too bland, salt actually below 1.5%     §SD-8         DISCARD           │
│  Brine fell, solids exposed              §SD-5         ADJUST (top up)   │
│  Brine very low + surface mold           §SD-5 → §SD-1 DISCARD           │
│  Too fast, mushy + over-acidic day 5     §SD-4         REFRIGERATE       │
│  Shio-koji clear at day 10               §SD-3 lower   ADJUST (warmer)   │
│  Koji ferment discolored                 §SD-3         DISCARD           │
│  Miso surface mold + paste below clean   §SD-1 miso    ADJUST (cap 1 cm) │
│      ↑ named structural exception (Doctrine v1.1)                        │
│  Miso paste below salt cap off / colored §SD-1         DISCARD whole tub │
│  Nuka surface mold + bed below clean     §SD-1 nuka    ADJUST (scoop)    │
│      ↑ named structural exception (Doctrine v1.1)                        │
│  Nuka bed off-smell from bed itself      §SD-1 + §SD-7 DISCARD whole bed │
│  Nuka bed unstirred 2+ weeks             §SD-7         DISCARD (default) │
│  Kombucha SCOBY moldy                    §SD-1         DISCARD SCOBY+jar │
│  Kombucha very vinegar (past day 14)     §SD-4         REFRIGERATE       │
│  Any ambiguity not in this table         (default)     DISCARD           │
└──────────────────────────────────────────────────────────────────────────┘

Default for any ambiguity: DISCARD. The §SD-1 asymmetry of cost (Doctrine v1.1) applies wherever this table does not.

A.6 — Glossary

Compact definitions of the technical terms the book uses repeatedly. Cross-references in parentheses point to where the term is introduced or used in depth.

Acid pickle. A vegetable preserved by added vinegar (acetic acid), not by microbial fermentation. NOT a ferment. Refrigerated only; 2–3 week shelf life. (§3.4)

Acidification. The drop in pH of a ferment as LAB produce lactic acid. A correctly-finished vegetable ferment reaches pH 3.4–3.8. (§4.5, §SD-2)

Aspergillus oryzae. The mold that is the koji culture. Cultivated for its enzymes (amylases break down starches; proteases break down proteins), not for acid production. (§3.2)

Brine. The salty liquid surrounding the solids in a vegetable lacto-ferment. May be the cook's own preparation (wet brine) or drawn from the vegetable by salt (dry-salt brine, as in sauerkraut). (§4.5)

Decision date. The cook's chosen point for moving a ferment from active room-temperature fermentation to refrigeration. Not the ferment's "completion"; the cook decides. (§SD-4, §4.7)

Discard. To throw away an entire jar or bed and start over. The default action on any ambiguous reading. The book uses "discard" consistently; not "throw away," "toss," or "chuck." (§5.1, §SD-1)

Kahm yeast. A flat-white surface film on a vegetable ferment, with no fuzz and no color. Distinguished from mold (which has texture or color). Indicates the ferment is at the edge of its safe window but is not itself unsafe; skim and continue, or end the ferment by refrigerating. (§5.3, §SD-1 kahm exception)

Koji. The general English term for rice or bean substrate inoculated with Aspergillus oryzae. Used in shio-koji, miso, sake, and several other Japanese ferments. Working temperature 22–35 °C (§SD-3), distinct from the 18–22 °C vegetable-ferment window.

Lacto-ferment / lacto-fermentation. A ferment driven by lactic acid bacteria. Examples: sauerkraut, kimchi, nukazuke. The acid is produced by the microbes, not added.

LAB (lactic acid bacteria). The bacterial group responsible for lacto-fermentation. Includes Leuconostoc mesenteroides (early phase) and Lactobacillus plantarum (later phase). (§1.2, §3.1)

Mold. Any visible colored or fuzzy growth on the surface of a ferment. Discard signal (§SD-1) except for the two named structural exceptions (miso salt cap, nukazuke bed; Doctrine v1.1). Distinct from kahm yeast (flat-white, no fuzz).

Mother of vinegar. A cellulose mat developed by acetic-acid bacteria in vinegar (and in kombucha). Sometimes confused with a kombucha SCOBY; the two are different organisms producing similar-looking films.

Nuka. Toasted rice bran. The substrate for a nukazuke bed. (§3.6)

Nukazuke. A Japanese pickle bed of fermented rice bran. Continuous-process ferment (the bed itself is the ferment); requires daily stirring (§SD-7). One of the two named structural exceptions to §SD-1 (Doctrine v1.1).

pH. A measure of acidity. The §SD-2 botulism boundary is pH 4.6; ferments correctly below this threshold are safe from Clostridium botulinum growth. A correctly-fermented sauerkraut reaches pH 3.4–3.8.

Salt cap. A layer of pure salt placed on top of a packed miso to seal the surface from oxygen and mold. Part of the structural reason for the §SD-1 miso exception (Doctrine v1.1). (§3.3, §5.2)

SCOBY. Symbiotic Culture Of Bacteria and Yeast. The cellulose mat hosting Acetobacter bacteria and yeasts that ferments sweetened tea into kombucha. (§3.5)

Shio-koji. A Japanese koji-based marinade: rice koji + salt + water, fermented at 25–28 °C for about a week. Enzymatic, not lacto-fermented. (§3.2)

Tamari (pooling on miso). The liquid that forms on top of an ageing miso as soy proteins hydrolyze. Not a failure; stir back in or reserve as a strong seasoning in itself. (§3.3, §5.12)

Vinegar pickle. See "Acid pickle." Used interchangeably in this book to emphasize the boundary case. (§3.4)

A.7 — Safety source notes

The rules in this book are the author's own teaching synthesis, but the underlying safety claims are well-established in published home-food-preservation and public-health literature. The cook who wants to verify a specific rule against an authoritative source can use the entries below as starting points. Each entry is intentionally short; this is not an academic reference list, and the cook is expected to consult the source itself if a question matters.

Salt is structural, not seasoning

The 2 % salt floor for vegetable lacto-fermentation, and the more general principle that salt concentration is the variable that selects which microbes can compete in the brine, are documented across home-preservation literature. The National Center for Home Food Preservation (NCHFP, University of Georgia Cooperative Extension) discusses fermentation as a salt-managed process and explicitly cautions against reducing salt levels in fermented vegetables for health reasons — salt is functionally structural, not flavor. The Notebook's §SD-8 floor (1.5 %) and recommended 2 % working level for vegetable ferments are within the NCHFP-supported range.

• Source: NCHFP — General Information on Fermenting (nchfp.uga.edu). For sauerkraut specifically: NCHFP's sauerkraut recipe is reference reading on the structural role of salt.

Garlic / herb / chili in oil is not fermentation, and is a botulism risk at room temperature

The aromatic-oil clarification in §SD-2 (Doctrine v1.1) — that garlic-in-oil, herb-in-oil, and chili-in-oil are NOT fermentations and must be refrigerated below 4 °C and used within four days — is grounded in public-health guidance documenting the Clostridium botulinum risk in low-acid foods stored anaerobically at room temperature.

• Source: Oregon State University Extension Service — Preserving Garlic and Herbs in Oil and related guidance. OSU advises that vegetables, garlic, and herbs in oil should be refrigerated below 4 °C and used within four days, or frozen. Storage of fresh garlic in oil at room temperature is not safe.
• Source: U.S. FDA guidance treats acidified low-acid foods (including garlic-in-oil mixtures) under 21 CFR 114; commercial production requires acidification or thermal processing to controlled standards.

These sources are the basis for the Notebook's hard line: garlic/herb/chili in oil should never be treated as fermentation, regardless of marketing language.

Acidified vs low-acid boundary (pH 4.6)

The §SD-2 boundary at pH 4.6 — above which Clostridium botulinum can grow in anaerobic conditions — is the standard public-health threshold used in U.S. food-safety regulation and home-preservation literature.

• Source: FDA — Acidified Foods (21 CFR Part 114). The pH 4.6 threshold defines the regulatory boundary between acidified and low-acid foods.
• Source: NCHFP — Acid Foods and Acid Levels discusses the pH 4.6 boundary and the safety reasoning.

Mold and discard-first

The §SD-1 discard-first rule for surface mold on home ferments is consistent with home-preservation literature, though the Notebook's articulation of the rule — and the named structural exceptions for miso and nukazuke (Doctrine v1.1) — is the author's editorial framing, derived from observed Japanese practice.

• Source: USDA / NCHFP — Mold on Food: Is It Dangerous? Surface mold is documented as a discard signal in most cases, with limited exceptions for hard cheeses and similar dense foods that do not apply to home ferments.

Temperature windows: this book's manageable windows, not absolute safety limits

This is worth saying directly. The temperature windows named in §A.3 (and throughout the book) are the windows the author has found reliable in a home kitchen across seasons. They are not the absolute lower or upper bounds named in published research. Many vegetable ferments succeed at temperatures below 18 °C (slower; sometimes cleaner-tasting). Some succeed at temperatures above 22 °C if the cook accepts more variability. The book's "working window" is what the author recommends for first-time fermenters and for repeatability across kitchens.

Conversely, the upper bounds in this book (e.g., > 28 °C as the vegetable-ferment failure edge) are conservative. In a perfectly clean kitchen with experienced technique, a sauerkraut at 26 °C may still succeed. The Notebook errs toward conservative windows for safety. The cook with more experience may push the upper bound; the cook should not push it without knowing what changed.

• Source: NCHFP — General Information on Fermenting discusses temperature ranges and the trade-offs between speed and reliability. The 18–22 °C working window in this Notebook is within the range NCHFP describes as reliable for home sauerkraut and lacto-ferments.

This book is not a regulatory document

A final framing note. The rules in this Notebook are practical safety guidance for home cooks. They are not regulatory standards, and the book is not a substitute for jurisdiction-specific public-health guidance. Cooks who plan to sell or share fermented foods commercially should consult their local food-safety regulator (FDA in the U.S., MHLW in Japan, EFSA in the EU) for production-scale requirements. Cooks with specific health conditions (pregnancy, immune compromise, severe allergies) should consult their medical providers before introducing new fermented foods to their diet.

The Notebook's claim is narrower: that for a home cook, following the rules in Doctrine v1.1 and the operational tests in Chapters 3–6, the six foundation ferments are reliably safe to make and eat. Beyond that scope, the cook is responsible for their own judgment.

Cross-references back into the book

The Appendix is short by design. When the cook needs depth, the body chapters are where it is:

• For why salt percentage matters: Ch.1 §1.2 (salt as threshold variable)
• For why temperature windows differ between ferments: Ch.1 §1.3 + §SD-3
• For what to do when something is wrong: Ch.5 in full
• For the cook's relationship with continuous ferments: Ch.6
• For the canonical safety rules and the named structural exceptions: Doctrine v1.1 (safety-doctrine-v1.md)
• For cross-chapter safety vocabulary: safety-language-canon-v1.md

The Appendix duplicates only what the cook needs at the kitchen counter. Everything else lives in the chapter that introduced it.

From the Notebook — Appendix closing note

Three uses for this Appendix the cook will return to:

1. Before starting a new ferment: read §A.2 (ratios) and §A.3 (temperatures) to confirm the targets.
2. During a ferment: print §A.4 (observation log) and §A.5 (decision table) and clip them near the jars.
3. When something looks wrong: use §A.5 first to make the decision; turn to Ch.5 if the entry is unfamiliar; consult §A.7 if the cook wants to verify the rule against an authoritative source.

The body of the book teaches the practice. The Appendix exists so the practice does not have to be re-read every time.