For most of my first year I treated starter liquid as packaging. The pellicle was the culture, the liquid was whatever came along with it, and if a friend wanted a SCOBY I fished one out with a slotted spoon and handed it over in a little cup of tea. Two of those gifts went moldy in a week. Mine never did, and the only thing I was doing differently was being lazy about draining the jar — I always left an inch or two of finished kombucha in the bottom because I could not be bothered to pour it all out. That laziness was the entire safety margin, and I had no idea.
The liquid is the culture, and the pellicle is luggage
The rubbery disc on top is mostly cellulose. Bacteria built it, and organisms live in it, but the working population of yeasts and acetic acid bacteria is dispersed through the liquid underneath. You can start a batch with liquid and no disc at all and it will build a new one. Start with a rinsed disc and plain sweet tea and you have handed a slow-growing culture an open, sweet, nearly neutral jar and asked it to outrun whatever mold spores are drifting through your kitchen.
That race is the whole problem. Sweet tea sits somewhere near pH 5, which is a comfortable place for things you do not want. Mature kombucha sits down near 2.5 to 3. Pouring enough of the mature stuff into a new jar drops the whole thing into acid territory on day zero, before the culture has done a single hour of work. Everything else — tea choice, sugar, temperature — decides how the batch tastes. The starter decides whether the batch is defended while it gets there.
How much I pour, and where the number comes from
Utah State University Extension's home kombucha instructions put it in kitchen units rather than percentages: reserve two cups of finished kombucha along with the pellicle to carry into the next ferment. In a gallon jar that lands right around one part mature liquid to seven parts sweet tea, and it matches what most published guidance works out to once you convert the units.
I run a little heavier than that. My gallon jars get a full pint and a half of finished kombucha, closer to one part in five, because my kitchen sits cool for half the year and a cool jar is a slow jar. Slow is fine. Slow and under-acidified is not. When the room is genuinely warm I go back toward the two-cup figure, since a fast jar needs less of a head start and too much starter races the batch past pleasant and into sharp vinegar before I get around to tasting it.
There is a ceiling, and it is a flavor ceiling rather than a safety one. Half and half gives you a jar that is unpleasantly sour on day four and never really develops. If you find yourself wanting that much starter to feel safe, the honest fix is a warmer spot or a smaller batch, not a bigger pour.
Measuring beats arithmetic, because pH does not average
The tempting move is to do the math: mature liquid at 2.8, sweet tea at 5, one part in seven, therefore the jar starts at some number you can work out on paper. It does not work that way. pH is a logarithmic scale and the mixture has buffering capacity of its own, so the blended number is not the weighted average of the two inputs and I have been wrong in both directions guessing at it.
So I test. A strip goes into the jar right after I have stirred the starter in, before the lid cloth goes on. The USU instructions ask for a reading below 4.5 at that point, with a tablespoon of plain white vinegar added and a retest if the jar comes in above it. That is the check I have run on every batch for three years and it takes about fifteen seconds. The finished-jar number I care about is tighter, but the start-of-batch reading is the one that tells me whether the batch was ever protected in the first place.
Where in the jar to draw it from
Take the starter from the bottom third, not from the clear liquid just under the pellicle. The yeast fraction settles, and the murky layer with the brown strands hanging in it is the part carrying the most working population. Those strands look alarming the first time and they are exactly what you want.
Draw it at the moment you bottle, not later. Kombucha left open on the counter for another week is still acidic but has burned through its sugar, and the culture in it is sluggish. If I know I will be a few days behind, I move the reserved liquid into a lidded jar in a cupboard with a spoonful of sugar dissolved in, which is the same logic my hotel jar runs on.
One hard rule: never take starter from a bottle that went through a second ferment. Flavored kombucha has fruit, juice, or ginger in it, all of which arrive with their own wild yeasts, and a batch started from raspberry kombucha is a batch you are no longer in control of. The plain first-ferment jar is the only donor.
Starting from nothing, and what a store bottle can do
If you have no mature liquid at all, a bottle of plain unflavored raw kombucha off a shop shelf works as the seed for a first batch. Buy the most boring one on the shelf, unpasteurized, no fruit, no added juice, and use the whole bottle in a half-gallon of sweet tea rather than a full gallon. A shop bottle has been refrigerated for weeks and the population in it is sleepy, so give it a smaller volume to conquer and expect the first cycle to take two or three weeks with nothing much happening on top for the first ten days.
That first batch is not a drink. It is a starter factory. It usually tastes thin and slightly off, and all of it goes into the second batch, which is the one that comes out right.
What a shortage actually looks like a week in
An under-started jar has a signature. Nothing forms on top for the first six or seven days where a normally started jar shows a skin by day four. The smell stays sweet and tea-like instead of turning sharp. And the surface is the place trouble shows first, which is why a slow start and a fuzzy spot so often turn up in the same batch.
If I catch it in the first two or three days I add more mature liquid and retest. After that I leave it alone and just watch, because opening and stirring a vulnerable jar repeatedly is its own risk. And if fuzz appears, the batch is finished — there is no version of adding acid after the fact that makes a moldy jar drinkable.
The two cups I never pour down the drain
Bottling day now has a fixed order. Pellicle out onto a clean plate, ladle two cups from the bottom of the jar into a measuring cup and set it aside first, bottle everything else, then build the new sweet tea and let it cool. The reserved liquid goes in last, gets a stir, gets a strip, and only then does the pellicle go back on top and the cloth go over the mouth.
Setting the starter aside before bottling rather than after is the single change that fixed this for me. When I bottled first, I would get to the end, find I had been generous with the last bottle, and start the next batch on whatever was left in the bottom — which is exactly how the two SCOBYs I gave away were packed. These days the gift is a full jar with a pint of liquid in it and written instructions to use all of it, and none of those have gone moldy either.