Fourth attempt. The first three are in the same notebook and they are where the temperature control came from: batch one dried into a crust with raw grain underneath, batch two hit 44C somewhere between two and five in the morning and came out smelling of ammonia, batch three was patchy because I eyeballed the spore dose. Total barley binned: 4.2 kg.
Barley rather than rice because barley koji, mugi-koji, is the traditional base for barley miso and barley shochu in Kyushu, and because pearl barley costs a third of what short grain rice costs in Malmö. It brings a strong protease and a decent alpha-amylase, so it handles protein and starch both.
Nothing clever. A 40 litre plastic storage box on its side, a seedling heat mat under a folded towel, a plug-in temperature controller with the probe pushed into the grain bed rather than taped to the wall of the box, a cheap hygrometer, a perforated stainless tray on two chopsticks so air gets underneath. Damp cotton cloth over the grain. Target ambient 30C.
Aspergillus oryzae wants 28 to 32C. Below about 25C the mycelium stalls. Above 40C it dies. Industrial koji rooms run 30 to 35C at around 90% relative humidity for roughly 48 hours, and the craft of it is that the mould heats its own bed as it grows, so the ambient number you set on hour one is not the number the grain sees on hour thirty.
If you are following along:
Read this before you buy anything. Koji temperatures are not a safety barrier: 30C is a fine growing temperature for plenty of things you do not want. The safety here comes from three other places. The grain is steamed. Steamed grain has low surface water, so it is a poor home for Bacillus cereus. And a heavy dose of A. oryzae takes the substrate before anything else gets a look in. This is a different arrangement from a fish ferment, where salt is doing every bit of the work on its own: @ferment/i-made-garum-for-six-months. So buy spores from a supplier who sells them for food, and never culture a mould you found on something. The food strains of A. oryzae cannot make aflatoxin, because the pathway is broken (mutations in the regulator aflR, deletions in the cluster), and it is GRAS in the US. Its wild relative A. flavus has no such defect, and ITS sequencing cannot tell the two apart, which is exactly why you buy rather than guess. Keep the bed under 40C. If a batch smells of ammonia, goes slimy, or grows anything black or blue-green, bin it. Do not sniff hard at it and do not taste it to check.
1.5 kg pearl barley into cold water, well covered, on the bench. Kitchen at 20C.
Drained into a colander for 30 minutes. The grain has gone from hard and glassy to matte, and it squeaks when you push a hand into it.
Into the steamer, 45 minutes. Barley takes longer than rice. Test is a grain squashed flat between thumb and finger: it should flatten without a hard white core and without smearing.
Spread on a tray to cool, fan on low across it. This is where you lose surface water, which you want: sticky grain clumps, and clumped grain grows unevenly.
Bed down to 34C. Spores weighed and mixed into a tablespoon of toasted barley flour first, so I could see where the dose was going, then rubbed through the grain with clean hands for three minutes.
Into the tray, bed about 4 cm deep, damp cloth over. Ambient 30.2C, bed 32.1C, RH 88%.
Ambient 30.0C, bed 30.4C, RH 90%. Nothing to see. Smells of wet grain and very faintly of the steamer.
Ambient 30.1C, bed 31.0C, RH 91%. The first white is visible if you push the cloth back and look across the surface at a low angle rather than down at it. Not fuzz yet, more of a haze.
Ambient 30.0C, bed 34.8C, RH 92%. It has started to make its own heat. Grains beginning to stick to each other. Smell now clearly something: sweet, a bit like a wet paper bag, and under that something raw and nutty.
Ambient 30.1C, bed 38.4C, RH 92%.
Broke the mat up with both hands, which takes effort by now, because the tray has knitted into one slab and it pulls apart in sheets. Spread thinner, raked three furrows the length of the tray with the side of my hand, cloth back on. Bed dropped to 33.6C within ten minutes.
This is the intervention that matters. The rule I work to now: past about 40 to 42C you mix immediately, and if it climbs back to 42C within two hours of mixing, the mould is making more heat than the bed can shed and you have to split it into two trays. Batch two died because I had no alarm set for four in the morning. On the subject of being awake at this hour, @nightshift/marcus-aurelius-at-four-in-the-morning is better than anything I could write about it, and I will only add that koji does not care either way.
Ambient 30.0C, bed 36.9C, RH 90%. Visibly white, properly furry between the grains. Smell has turned: chestnut, mushroom, a hint of cut grass.
Ambient 30.0C, bed 41.2C, RH 89%. Mixed and split it into two trays at about 2.5 cm deep, and dropped the ambient setpoint from 30C to 27C. Bed settled at 34.1C and stayed under 38C for the rest of the run.
Ambient 27.1C, bed 36.2C, RH 88%. White and dense, and the trays now smell strong enough to notice from the hall. Sweet and floral over the chestnut.
Ambient 27.0C, bed 35.4C, RH 87%. Faintly yellow-green in two patches where the bed is deepest, which is the mould starting to make conidia. Mature A. oryzae spores are mustard-yellow to olive-green and that is normal for the species: black koji is a different organism, A. luchuensis, used for shochu and awamori. For cooking koji you want it before heavy sporulation, because the spores bring a musty edge, so this is the signal that the clock is running out.
Ambient 27.0C, bed 33.8C. Out.
2.9 kg of finished koji from 1.5 kg of dry barley. One white slab per tray, holds together when lifted, breaks with a sound like a biscuit breaking. Smells sweet and mushroomy. Raw, a grain is faintly sweet at the front and chalky at the back.
Shio koji: 600 g of the barley koji broken up, plus 900 g water, plus 225 g salt, which puts the salt at 13% of the finished 1725 g. Jar loosely lidded, on the bench at about 20C, stirred once a day for eight days, then into the fridge. I use 13 rather than the 10% a lot of recipes give, because commercial shio koji sits around 13 and below roughly 12 you are inviting company. It keeps about six months cold, and half a teaspoon rubbed into chicken thighs two hours ahead does more than an hour of marinade.
The remaining 2.3 kg went into a barley miso with cooked soybeans, salted to 12% of total weight, weighted, into the cellar at 14C. That gets opened next autumn and will be its own notes.
Net across four batches: 4.2 kg of barley binned, 2.9 kg of koji made, and one rule learned, which is that the ambient number is a suggestion and the bed number is the truth.