Yellow-koji starter

What the starter is

Yellow-koji starter is seed in the most literal brewing sense: the spores of the yellow koji mold, grown out on cereal and then harvested, dried and bagged so that a brewer or a home cook can plant them on steamed rice and start again. Japanese specialist suppliers have sold it for centuries under the name tane-kōji, 種麹 or 種麴, and older trade usage also called it moyashi, もやし, or tane-moyashi, 種もやし. The National Research Institute of Brewing's English glossary lists the same word under four headwords — koji spores, seed koji, starter koji and spores of koji mold — and keeps it well clear of shubo or moto, the sake starter culture, which is a yeast-and-bacteria mash rather than a mold inoculum. It is not the same thing as koji either. Koji is grain with the mold actively growing on it, alive and full of enzymes; the starter is the inoculum that makes koji, and the two have different jobs, different cultivation times and different shelf lives.

The word "yellow" does the work of a species label. Japanese brewing taxonomy sorts the koji molds by the colour of the spores each one produces, and the yellow group covers Aspergillus oryzae, the species behind sake, and Aspergillus sojae, the species favoured for soy sauce, along with albino mutants of both. Those spores read yellow-green to green, with the shade shifting according to the medium they were grown on and how long they were left to sporulate. The colour is literally the name: the institute notes that the black in black koji mold and the yellow in yellow koji mold come from the spore colour itself. The other branches of the family are excluded here — Aspergillus luchuensis, including var. awamori and mut. kawachii, which carries awamori and most shochu, and A. niger, which brewing does not use at all.

A starter packet is therefore a small, dry, deliberately boring object: pale powder or pale grains, faintly earthy, with no sweetness or savoury punch of its own. Everything interesting happens downstream. A. oryzae and A. sojae were designated part of Japan's national fungus, kokukin, by the Brewing Society of Japan in 2006, a designation given to the collective koji-kin rather than to any one packet, and the mold's identity has a documented Western paper trail too: Hermann Ahlburg, a foreign advisor in Meiji Japan, isolated it from rice koji in 1876 and named it Eurotium oryzae before it was renamed A. oryzae. The collective term koji-kin itself only came into general use in the Showa era.

Three things are easy to confuse and worth separating before going further. The first is the mold, a living fungus. The second is the starter, the harvested spores sold to inoculate grain. The third is koji, the grain product those spores make. A fourth entry, the grown koji sold in packets for home miso or amazake, is a preparation rather than an ingredient, and it is not this page. This record covers the yellow-koji inoculum as bought or kept: the granular and powdered spores that a rice-koji recipe calls its seed.

From wood ash to a national trade

Before there was starter, there was luck. The oldest way to make koji was to wait for mold to arrive on steamed grain, or to carry a little of a good batch into the next one. Both were unreliable, and the carrying method left a measurable trace: when researchers examined the in-house starters that breweries still kept under the tomo-dane practice, Penicillium turned up in five of seven companies' products, bacterial and yeast counts ran from 8,000 to 300 million per gram, and the ratio of bacteria to spores sat between 0.23 and 2 percent. The eighth- to tenth-century Engishiki prescribes inoculating with about a tenth of the koji weight, roughly a hundred times the dose a modern brewer uses, which says less about old recipes than about how weak the inoculum was.

The turn came with ash. Practitioner accounts date the mixing of wood ash into a koji culture to around six hundred years ago, in the Muromachi period, and describe it as the discovery that made seed koji possible: ash raises the alkalinity of the substrate, keeps the mass open so oxygen reaches the mycelium, and supplies minerals that push the mold to form durable, dispersible spores rather than just spreading vegetatively. The ash was not swept from a hearth but made to order, steamed and roasted from the branch tips and leaves of camellia and oak, and the starter houses bought it from specialist ash producers, the hai-ya. A related account places the wood-ash technique in the Muromachi period as well and treats it as the start of tane-koji as a distinct product.

Once spores could be kept, they could be sold. Kōjiya Sanzabemon, founded in Kyoto in the Muromachi period, was granted permission to sell seed koji by the Ashikaga shogun Yoshiteru, who ruled from 1546 to 1565 — a licence to trade a microorganism, six centuries before the phrase "culture supplier" existed. The same firm still makes starter today; it moved to Toyohashi in Aichi in 1965 and set up a research company, B'ioc, in 1992. By the middle of the Edo period seed koji had become a separate craft with its own cultivation methods, distinct from koji making, and the manuscript of one Kyoto house records customers ordering specific characters — a "strong moyashi", a "white koji" — and the house testing rice varieties for how well they carried seed koji. Kyoto dealers named in the records include Kōjiya Sanzabemon, Omiya Kichizaemon and Hishiya Rokuzaemon, with Osaka houses behind them, and buyers came from Michinoku in the north to Hizen in the south. An older account of the koji trade adds that koji-making specialists, the koji-za, had monopolised the craft before the Bun'an koji disturbance broke their hold, after which two tane-koji dealers are recorded in early Edo Kyoto.

The modern commercial shape dates from much later. If "for sale" is written into the definition, the history of koji starter as a product begins in the Meiji era, 1868 to 1912, sold as a naturally dried granular product made mostly of grain; records of it being used for miso follow in the 1900s. Sake competition spread the habit of buying starter to miso and soy sauce houses, and the number of makers grew — around ten from Meiji, still fifteen in the Showa era by one 1984 count — before consolidating again. The National Koji Starter Association had 43 member companies in 1949; today twelve remain, of which only six sell nationwide, a figure the head of one starter house gives as six national makers plus two serving local customers only. Manufacturing and selling a microbial seed, rather than the fermented food itself, has been described as an industry unique to Japan, and as the first biotechnology industry there: it produces the microorganism separately from the product.

Purity arrived in stages. The importance of using pure original strains was proposed in 1895 and a method for culturing them and making starter from them was reported in 1911, but it became mainstream only after the Showa era began; large soy sauce makers led the shift to pure isolation from the late Meiji through the Taisho era, and designing a starter around a strain's known characteristics became general practice only after the Second World War. In 1907 three strains were first isolated from sake koji starter, and in 1926 it was proved that starter contributes koji mold alone, not the bacteria that had been riding along with it. The hygiene gains are legible in the numbers: bacterial counts in starter ran from one million to one hundred million per gram in the early 1900s, fell sharply after about 1960, and by the 1990s most powdered products sat under a thousand per gram. A brewer's decision to buy rather than propagate starter is now near universal, and the reason offered is practical rather than romantic: koji mold grows only under conditions that are hard to create and hold.

How a starter is made, and why it is not just long koji

Starter making is an extension of koji making, but the aim is inverted. Brewing koji is grown for enzymes: the mold is encouraged to penetrate the grain and load it with amylases and proteases, and the job is finished in 40 to 45 hours. Starter is grown for spores, with as few bacteria as possible and enough durability to sit on a shelf, and that takes four to six days, about three times as long, depending on the strain. The same open-room craft therefore splits into two different calendars, and the difference is visible in every step.

The substrate is treated differently from brewing rice. Starter grain is polished only lightly, 2 to 3 percent for rice and 10 to 20 percent for barley, leaving the bran rather than stripping it, and the material is brought to around 40 percent moisture. Wood ash, 0.1 to 0.5 percent, is dusted on as the steamed grain cools, or at inoculation — the single step that most clearly separates starter work from brewing work. The pure strain is then sprayed on at a grain temperature of about 30 to 32 °C, in a room kept near saturation so the substrate does not dry out, while the mold's own growth optimum sits higher at 35 to 38 °C. The mass is turned once, at the teire stage around 20 to 24 hours in, and then deliberately left alone: no further stirring means the mold stops spreading through the grain and starts throwing up conidia on the surface. Late in the culture the temperature is held near 30 °C. Grain temperature across the whole run typically sits between 30 and 35 °C, with the pattern and the peak varying by strain.

Harvest is a sorting operation. The grown grain is sieved through 100 to 300 mesh to release the spores, and each batch is checked for appearance and enzyme activity to confirm the strain has not drifted during production. Then it is dried, below 40 °C, to 10 percent moisture or less; a food-safety guideline puts the industrial drying at 40 to 42 °C in dehumidified air, with the dried spores held at 5 to 15 °C afterwards. That drying step is what turns a perishable mold culture into a tradeable good, and it is the modern descendant of the oldest storage trick in the trade: at the end of the brewing season, lime and wood ash mixed in equal parts were placed in wooden boxes with the starter, sealed and kept dry for the next season, a practice that carried on until the late 1800s.

Two products come out of this. Granular starter, 粒状, is the dried grain bagged as it is, spores still sitting on the rice or barley it was grown on; it is what handmade koji uses, and shochu makers still buy a high proportion of it. Powdered starter, 粉状, has the spores sieved off the carrier and the count standardised with alphalised starch as a bulking agent; it suits mechanical inoculation devices, and it now makes up about 90 percent of what is sold. Nearly all miso and soy sauce brewing uses powder, as do most large sake breweries; the granular form survives where the work is done by hand.

The spores themselves are about five micrometers across, and a gram of finished starter carries anywhere from five to thirty billion of them depending on the strain. That range is why inoculation is quoted as a proportion rather than a fixed weight: roughly one part in a thousand of the material for granular starter, and one part in three to six thousand for powder, which works out to something like 5,000 to 10,000 spores landing on each grain of steamed rice. The British Columbia guideline that treats koji making as a controlled process asks for a certificate of assurance on purchased Aspergillus strains, with spore counts above about one billion per gram and aerobic contaminants under 1,000 CFU per gram — a floor well below the commercial norm, and a reminder that the useful measure of a starter is not only how many spores it has but how little else came along with them.

Strains, colours and what each one is bred for

Only three Aspergillus species are used in Japanese brewing — A. oryzae, A. sojae and A. luchuensis — but around forty strains circulate among starter makers, and many commercial starters are blends, put together to soften one strain's excesses or to combine two strains' strengths. A starter is thus closer to a seed-catalogue variety than to a single organism: the species tells you the family, the strain tells you what the koji will do.

The trade sorts its catalogue three ways at once. By use: starters for sake, subdivided into daiginjo work, moromi work, aromatic strains and liquefying mashes; for miso, split by rice, barley and soybean miso and further by red, white and Saikyō styles; for soy sauce, with koikuchi, usukuchi, soya and white soy sauce strains; for shochu, rice, barley and awamori; and for other work such as amazake and the sale of fresh koji. By form: granular or powder. By colour: yellow, white and black. The colour axis is where the yellow-koji category has an interesting wrinkle, because albino mutants of A. oryzae and A. sojae are still yellow-koji molds in the taxonomic sense even though their colonies read white. A record from the Bunka years, 1804 to 1818, already notices the shift, remarking that koji "has always produced yellow flowers" — meaning spores — but that white koji without spores had recently come to be considered superior. The white mutant of black koji, isolated in 1918, is the famous case in shochu; the white mutants of A. oryzae are the quiet ones, widely used in miso and increasingly sold to home makers because they keep amazake and white miso pale.

Enzyme balance is the real choice. The National Research Institute of Brewing explains that koji supplies the enzymes that break down starch and protein along with the flavour and aroma compounds of the finished food, and because each product wants a different balance, dedicated strains are used — a sake koji mold, a miso koji mold. A European culture vendor selling nine strains puts the same idea in practical terms: a light rice strain with strong starch-splitting power for short, sweet miso, amazake and sake; a red rice strain bred for long, saltier rice miso with more protein breakdown and less saccharification; a barley strain for barley miso; a soy strain for all-soy miso; and a soy sauce strain bred to leave sugars behind for the yeasts while working hard on protein. The vendor treats the miso strains as broadly interchangeable but insists on a proper soy sauce strain, and notes that a rice strain will grow on soybeans, just less well than on the substrate it was bred for. That guidance is a supplier's, not an institution's, and its species labels are inconsistent, so it is worth using for the logic of strain selection rather than for taxonomy.

Where the yellow molds stop matters too. Sake is made overwhelmingly with yellow koji mold, with red koji mold and Rhizopus as the documented exceptions. Shochu and awamori run on black or white koji mold, and the difference is chemical, not cosmetic: koji made with black or white molds carries large amounts of citric acid and supplies an acid-tolerant amylase, which acidifies the moto and the moromi and blocks harmful contamination. Yellow koji does not bring that acid shield, which is one reason sake brewing has evolved its own acidification steps. Some rice shochu, and breweries recreating Meiji-era methods, do use yellow koji mold, so the boundary is a documented tendency rather than a wall.

What the starter brings to the foods that need it

Yellow-koji starter is the input to every koji that carries the yellow mold, and those koji are the engine of a large part of Japanese cooking. Rice koji, the most familiar, is steamed rice planted with these spores and cultured for a day and a half; it is the saccharifying agent in sake, the starch source in amazake, the starter culture in shio koji and the grain component of rice miso. Barley koji does the same job for barley miso and barley shochu; soybean koji carries the mold into soybean miso; and the wheat-and-soy koji of soy sauce brewing is one of the largest single markets for powdered starter in the country. By volume the starter trade runs soy sauce first, then miso, then shochu, then sake — a ranking that surprises nobody who has priced a bottle of supermarket shoyu against a bottle of ginjo.

What the mold contributes is enzymatic, and the starter decides how much of each enzyme arrives. A. oryzae and A. sojae send glucoamylase and other amylases into the grain to free sugars from starch, and proteases to free amino acids from protein. Sake needs sugar for yeast to turn into alcohol, so its koji is bred and grown for saccharifying power. Miso needs both sugar, which feeds the lactic acid bacteria and yeasts that sweeten and perfume it, and amino acids, which supply savoury depth over months of aging; a long-aged, saltier miso rewards a strain with more proteolytic reach. Soy sauce needs protein above all, and needs the sugars left unspent so that the yeasts and the salt-tolerant bacteria can build its aroma. This is why a starter sold "for soy sauce" behaves differently from one sold "for amazake" even though both are yellow koji mold, and why a home cook who substitutes one for the other gets a workable but lopsided result rather than a failure.

The same spores also reach the home kitchen in a form that has almost no historical precedent. A Japanese supplier sells a home starter as a pure white strain named Yukikomachi, described as widely used for amazake, miso and processing, in a 70 g bag reckoned to cover 200 kg of raw rice; its worked example uses 1 to 2 g of starter per kilogram of steamed rice, sprinkled on once the rice has cooled to around 45 °C. That is a household instruction, not a validated safety process, but the arithmetic matches the industry's own dosing logic: a gram of powder holding billions of spores needs only a pinch per kilogram of grain to reach thousands of spores per grain. A European vendor sells the same class of product as spores with a separate note on diluting them before use, which is the practical way to spread a sub-gram dose evenly through a kilo of rice.

Buying, dosing and keeping it

Choose by form first. If the koji will be made by hand, in trays or a small box, granular starter is the natural match: the spores come already sitting on grain, it is easy to see and easy to spread, and the traditional dosing figure of about one part in a thousand of the material assumes that form. If the koji will be made in a machine, or if the work is done at home in small batches where an even distribution matters more than tradition, powder is the better buy, dosed at one part in three to six thousand and best diluted with a little flour or starch before sprinkling. Read the label for the species and the intended use, since a packet sold for soy sauce or for long-aged miso is bred for a different enzyme balance than one sold for amazake.

Dosing is a numbers game that rewards a rough approach. With a finished starter carrying five to thirty billion spores per gram, the industry's own targets of 5,000 to 10,000 spores per grain of steamed rice are met by weights measured in grams per kilogram, not by cups. A home example of 1 to 2 g per kilogram of rice sits in the same order of magnitude as the granular one-in-a-thousand rule. Too little starter leaves bare grains that wild molds and bacteria can colonise; too much does no great harm but wastes a scarce product and can push the mass to overgrow.

Keeping the starter dry is the whole of keeping it. The industrial pattern is drying to 10 percent moisture or less at under 40 °C, then storage between 5 and 15 °C, and the published result of that drying is a product that keeps at room temperature for six months or more as long as it does not pick up moisture again. Starter houses themselves keep their dried starter around 15 °C at 50 percent relative humidity, hold stocks in more than one place, and re-culture as little as possible so the strain does not drift. A home packet is best kept sealed, dry and cool; the sources read give no home shelf life beyond the six-months-or-more figure for properly dried starter, and no basis exists for inventing one. A refrigerated or frozen packet of fresh, wet koji — grown koji rather than dried spores — is treated by the safety guideline as good for about a month, or freezable, with slightly reduced activity acceptable as long as it has been protected from contamination.

Safety sits on both sides of this product. The mold itself is about as well-established a food microorganism as exists: safety was argued first from centuries of use in sake and miso, then demonstrated scientifically from the 1960s onward. After turkey deaths in England in 1960 were traced to aflatoxin in peanut meal contaminated by A. flavus, Japanese national and private bodies inspected the brewing molds thoroughly and confirmed that none of them produces aflatoxin; genome work in 2005 separated A. oryzae from A. flavus, and analysis of the aflatoxin gene cluster placed every starter strain examined in the non-aflatoxigenic groups. Japan's National Institute for Infectious Diseases classifies koji mold at biosafety level 1, meaning it is not expected to cause disease in healthy people, and the US FDA lists it on the GRAS list. The relationship to A. flavus is real, and it is the reason for the one rule that a home maker should not bend: the safety guideline advises against back-slopping, that is, against using one's own koji to produce a new batch of starter, because the risk of amplifying wild Aspergillus strains that do produce aflatoxins goes up sharply. Buying the spores and not trying to propagate them at home is the documented position.

The other hazard is not the mold but the ones that arrive uninvited. Koji is made in an open system, and the common intruders are Micrococcus, coliforms introduced by hands, and Bacillus, whose heat-resistant spores multiply quickly once the grain runs above 35 °C and survive even high salt. That is why both brewing koji and starter work begin at a cool grain temperature, why the BC guideline asks for cooked rice to be cooled to 35 °C within two hours of its internal temperature falling below 60 °C before spores go on, and why incubation is held around 30 °C and never allowed past 40 °C. Koji workers, the literature notes, traditionally avoid eating natto before a shift, since it is a dose of exactly the bacterium they are trying to keep out of their trays.

Notes and open questions

The dating of the trade depends on how strictly the word "sold" is read. The head of a Kyoto starter house dates the wood-ash technique and his own firm's licensed sale to the sixteenth century; the Journal of Fungi review grants that in-house starters, including the tomo-dane method, existed long before and places the history of starter "for sale" in the Meiji era. Both are recorded here, and the period on this page marks the earliest dated commercial attestation rather than an origin.

Whether A. oryzae was domesticated by humans or simply selected during koji making is still argued; the split from A. flavus is reported at around 1000 BC, roughly when rice cultivation reached Japan. No claim about domestication is made here.

Home-scale storage figures are thin. The published numbers are industrial — drying at 40 to 42 °C to 10 percent moisture, storage at 5 to 15 °C, and six months or more at room temperature for starter dried to 10 percent moisture or less — and no retail shelf life appears in the sources read. Carrier allergens come from the packet label: the granular form is grown on rice or barley, and powder grades are bulked with starch.

The European vendor's species labels conflict with Japanese brewing taxonomy in places, notably in calling a black-koji strain A. awamori where the brewing literature now uses A. luchuensis. Its strain advice is used here only for the logic of choosing by intended product.