Baumy Biltong
Seven batches. The first one cut to a spec rather than to the seam.
An 8.7 kg basis. Single-vinegar wash, MSG in the dredge, 25 mm cut along the grain, and a hot-start then cool-finish drying curve aimed at a defined bark over a red interior.
Why the seventh revision
Sixth revision → seventh revisionFirst revision cooked and measured rather than planned. Five real changes. (1) Wash collapsed from five vinegars to one: South African brown spirit vinegar, which is the traditional single vinegar, with Worcestershire kept for the glutamate and inosinate the anchovy brings. (2) MSG added at 20 g, roughly a fifth of salt weight, to lift a profile that reads flat once the salt drops. (3) Cut specified at 25 mm along the grain with silverskin removed, rather than whatever the seam gave. This is the largest change and it rewrote the timeline: see the drying rate note. (4) Toast sequence formalised into three pans, and the grind moved to mortar and pestle, which needs about 15% more volume for the same weight. (5) Pull target split by thickness rather than one figure for the batch: 46 to 52% for pieces over 20 mm, 50% for 12 to 20 mm, 55% under 12 mm. Salt raised to 15.5 g/kg to compensate for the wetter pull, since finished salt concentration is green salt divided by finished mass and a 46% pull leaves roughly 17% less salt in the product than the 55% pull of batches 4 and 5. Nitrite was specified and then dropped for lack of supply, so the deep red interior remains unsolved.
Prep
4 steps-
Trim the silverskin off. It is dense collagen and a vapour barrier: leave it on and that face will not dry while everything around it does. Cut along the grain, 25 mm on the short axis, roughly 40 mm wide. Sort what you get into three groups by thickness: A over 20 mm, B 12 to 20 mm, C under 12 mm. Weigh every piece and record it.
Moisture diffuses two to three times faster along muscle fibres than across them, so cutting with the grain puts the fast axis down the length of the strip and the slow axis across the 25 mm. The ends breathe and the long faces build bark.
Work Butchery -
Three pans, not one. The seeds differ in mass and moisture and will not finish together. Pan one, coriander alone, medium and dry, 3 to 4 minutes, shaking constantly: pull it the moment it smells citrus-floral and the first seeds pop, and do not chase colour. Pan two, black pepper and star anise together, low-medium, 4 to 5 minutes, until the pods fracture. Pan three, chillies deseeded and torn flat, lowest heat you can hold, 60 to 90 seconds a side pressed against the pan, off the moment they are pliable and smell of raisin and tobacco. All three onto a cold tray, spread thin, cooled to room temperature before grinding.
Overshoot the chillies by fifteen seconds and the whole batch goes acrid. Capsanthin and the residual sugars scorch well before the flesh looks dark.
Time 20 minWork Toasting -
Mortar and pestle, decanting between each. Star anise and pepper first, medium-fine, while the mortar is clean and dry. Chillies second, coarse flake, because they will slick the mortar with oil. Coriander last and in two loads, coarse and deliberately under-worked: cracked halves and quarters with visible seed structure, about a third of it barely broken. Stop early.
A mortar gives a wider particle distribution than a blade grinder and coarser fragments pack worse, so the same weight covers less. Allow about 15% more volume than a blender figure.
Work Grinding -
Into a bowl of at least 3 litres: the ground spices first, then the untoasted powders in order of bulk, tandoori and salt before MSG, cayenne and garam. Toss dry with your hands, fifteen turns, no oil and no wash. Mix within an hour of grinding and dredge within a few hours of mixing.
Cure
3 steps-
Combine the wash. Work fast and keep the meat cold: as it warms, fat and myoglobin bleed in and the wash stops being usable. Massage each piece firmly into the grain for two to three minutes. You are looking for visible surface tack before the dredge goes on.
The tack is myosin, not myoglobin. Salt above about 0.3 M ionic strength solubilises myofibrillar protein and agitation drags it to the surface, where it dries into a pellicle: a bound layer that forms cohesive bark rather than crumbling, and that holds the dredge on. Myoglobin is sarcoplasmic and leaches out passively no matter what your hands do.
-
Dredge to full coverage on group A, working heavy. Go deliberately lighter on group C, around 60 to 70% of that visual density with meat showing through. A 12 mm piece carries roughly twice the surface area per kilogram of a 25 mm piece, so coating both to the same density puts nearly twice the salt concentration on the thin one.
Dredge A first with a heavy hand and C last with a light one.
Work Dry-curing -
Rest the dredged pieces in the fridge at 3 to 4 degrees for 48 hours. Weigh each piece before it goes in: this is functionally a dry brine and the number you take here is what every later calculation measures against.
Time 2 dWork Dry-brining
Hang
3 steps-
Group A only. Hook and hang with the 25 mm face into the airstream, at least 40 mm between pieces so adjacent boundary layers do not merge into one stagnant pocket. Heating pad on, box air held at 32 to 34 degrees, fan at full, for 16 hours. Probe at meat level rather than trusting the pad setting; wood insulates and the two will disagree by ten degrees.
This is a controlled case hardening. Front-loading the heat sets a 2 to 4 mm crust early, which then restricts further loss and holds the steep gradient that makes a bark. Groups B and C stay in the fridge through this phase: bark depth is close to absolute rather than proportional, so 3 mm from each face of a 12 mm piece is half the cross-section and on a 10 mm piece it is the whole thing.
Time 16 hTemp 33 °CTool 7 cm steel hooksTool drying boxTool heating pad -
Pad off. Groups B and C go in now. Hold ambient, 24 to 26 degrees, fan running to the end. The fan must exhaust rather than recirculate: this batch evaporates around 6 kg of water, near 500 g a day in the first three days, and a sealed box pins at 100% relative humidity and stops drying entirely. Inlet low, outlet high. Thickest pieces nearest the fan, smallest in the deadest corner permanently. Rotate A and B at 72 hours and again at 7 days.
-
Weigh per piece against its own hanging weight, hook on, and pull by group. Group A at 46 to 52%, group B at 50%, group C at 55%. Group C goes furthest because there is no gradient to preserve: the whole cross-section sits at one water activity, so 46% leaves it soft rather than red and wet. Trust the scale, not the calendar.
Time 4 d–11 dWork Air-drying
Notes
630 × 7 cm steel hooks, 13 g each — €29.97. The 13 g is subtracted from every gross weight to get net.
Dried spice is a live microbial vector. Whole seeds and pre-ground powders both carry mould conidia, mostly Aspergillus and Penicillium, and pre-ground powder tends to carry more because grinding exposes interior surface and the material sits warm. Dredged onto wet meat that then hangs for two weeks at 25 to 34 degrees, that inoculum lands on an ideal substrate.
From this revision on, everything gets a thermal treatment, not only the whole seeds.
Whole seeds and chillies: the three-pan toast already in the method does this.
Powders, which are tandoori, garam, cayenne and MSG: these scorch in a pan, so spread them on a tray in a low oven at 110 degrees for 15 minutes instead, then cool fully before mixing. Gentle enough not to damage the paprika pigment or drive off what aromatics the powders have left.
Be honest about what this achieves. Dry heat at these temperatures knocks down vegetative cells and mould conidia substantially but does not sterilise, and it will not touch bacterial endospores, which need far higher temperatures or moist heat. Mould conidia are much less heat-resistant than bacterial spores, which is why the treatment is worth doing for this specific risk. It is a reduction in inoculum, not a guarantee.
Spice is also not the dominant vector. Ambient spores, surface moisture and dead airflow zones matter more. The acetic wash and adequate air movement remain the main controls. This closes one door that was standing open for free.
20 g across 8.7 kg green, which is about 0.23% green and 0.43% of finished weight at a 46% pull. Roughly a seventh of salt weight. It goes in untoasted with the other powders: MSG is non-volatile and stable well past 200 degrees, so there is nothing to gain from heating it and nothing lost in the dry.
The restraint is deliberate. Three glutamate sources already stack in this product before any is added. Fourteen days of endogenous cathepsin and calpain activity liberates free amino acids from the myofibrillar protein, glutamate prominently, which is the same mechanism behind dry-aged beef. Losing 46% of mass then roughly doubles the concentration of every non-volatile solute, so whatever the meat generates is eaten at double strength. And the Worcestershire is anchovy-based, bringing inosinate, which binds the T1R1 and T1R3 receptor cooperatively with glutamate rather than additively: up to around eightfold potentiation at favourable ratios.
So a modest addition against that background does considerably more work than the same addition alone. Above roughly 1% of finished weight, glutamate reads soapy and metallic. Doubling this would push past the receptor's useful range and make it taste of commercial spice rather than dried beef.
Sodium cost is trivial: MSG is 13.6% sodium against salt's 39.3%, so 20 g adds about 2.7 g of sodium across the whole batch.
The batch 2 model, 4.21% of initial weight lost per day and 13 to 15 days for a medium piece, does not apply to this revision and should not be used to plan it.
Measured on batch 7 at the 25 mm cut: pieces of 400 to 480 g net reached 42 to 49% loss in 8 days, which is 5.3 to 6.2% per day. Batch 2 ran about 2.1 to 2.7% per day on pieces of 500 to 700 g. Roughly double the rate.
The cause is thickness, not airflow or temperature. Internal moisture diffusion is the rate-limiting step once the surface has dried, and diffusion time scales with the square of the half-thickness. Halving the thickness quarters the diffusion time, and the surface drying rate is unchanged. Batch 2 was cut to whatever the seam gave, which was considerably thicker.
Plan group A at 9 to 11 days rather than 14 to 17.
Batch 7 mixed 486 g of dredge and applied roughly half of it. Adhesion works out at about 32 g per kg of meat, so a mix sized at 56 g/kg was close to double what the meat could take.
Two consequences.
First, sizing. Mix around 45 g/kg, which still leaves 40% headroom against the coverage problem that left 1.6 kg bare in batch 4, plus a separately ground reserve of about 200 ml. Do not size the mix by the salt target, because the salt target is only met if the whole mix goes on.
Second, and more important: dose salt by what adheres, not by what is mixed. Batch 7 specified 15.5 g/kg and delivered about 9 g/kg because half the bowl was left over. That is well below the 12.7 to 14.1 g/kg band of batches 2 to 5 and it is the single biggest error in the batch.
There is also a structural unevenness that cannot be dredged away. Spice pickup scales with surface area and meat mass does not, so a thin 60 g offcut carries close to 1.7 times the salt density per gram of a 500 g piece cut at 25 mm. In a mixed-thickness batch the small pieces will always be saltier. Judge seasoning on a large piece, never on an offcut.
Salt is denser and finer than tandoori or cracked coriander and it migrates downward as the bowl is tossed. If only part of the mix gets used, the salt figure that actually reached the meat depends on whether the dredging was done off the top or from the bottom, and it can be well off the nominal fraction.
Untested, but it makes every back-calculation from a partly-used mix uncertain by an unknown amount. Two ways to close it: weigh the leftover mix at the end and subtract, which gives the applied weight exactly but still not the applied composition. Or keep the salt out of the dredge entirely and apply it separately by weight per piece, which fixes composition but loses the even distribution the dredge gives.
Provenance
Started as a South African staple made from Irish silverside in Berlin, which is most of the reason the spice profile drifted toward tandoori masala.
The pattern
cured dried meatThe plain idea under this dish. Cook it with what you have. The full recipe is on the Method tab.
beef strips cut along the grain + vinegar and Worcestershire wash + salt cure + toasted coarse spice dredge + chilli + MSG, dried in moving air to a target weight loss
Other dishes with this pattern: cured dried meat.
The parts
6Beef silverside, silverskin removed, cut along the grain to 25 mm
The meat that dries. The 25 mm cut along the grain sets the drying time and lets a bark form on the long faces.
Wash: South African brown spirit vinegar and Worcestershire sauce (139 g vinegar and 141 g Worcestershire for 8.7 kg meat)
Wets the surface so the dredge sticks. The acetic acid is one of the main mould controls. The Worcestershire adds savoury depth.
- !SAFETY Swap: Four vinegars, as in revision 6: 45.2 g red wine, 36.7 g apple cider, 43.2 g condimento blanco and 34.6 g black malt vinegar for 10 kg meatKeep the total vinegar weight the same. Revision 6 used 159.7 g vinegar for 10 kg, which is about 16 g per kg. Revision 7 uses 139 g for 8.7 kg, which is also about 16 g per kg. Do not reduce the vinegar: the acetic wash is one of the main mould controls. Do not make the wash heavier: acid pushes an uncured interior toward brown. Revision 7 changed to one vinegar because brown spirit vinegar is the traditional single vinegar.
Salt, 135 g for 8.7 kg meat (15.5 g/kg green weight), mixed into the dredge
Draws water out and makes the meat hostile to spoilage before and during drying. Sets the final seasoning.
- !SAFETY Swap: Nitritpökelsalz (German nitrite curing salt) in place of the salt, weight for weightUse Nitritpökelsalz only. It is 0.4 to 0.5% sodium nitrite in salt. Replace the salt with it, weight for weight. Do not add it to the salt. The science note calculated 13.85 g/kg, which gives about 69 mg/kg nitrite. Revision 7 uses 15.5 g/kg, and the note did not calculate that rate. The note adds sodium ascorbate at 0.5 g/kg. Nobody has cooked this dose here yet. Do not use Prague powder #1 for this swap. It is a much stronger cure, and the same weight gives many times the nitrite.
- !SAFETY Swap: Salt at 13.85 g/kg green weight, the rate of batches 4 to 6Use this rate only with the older pull target of 55% weight loss. Revision 7 raised salt to 15.5 g/kg because it pulls wetter, at 46%. A 46% pull leaves roughly 17% less salt in the product than a 55% pull. Weigh the salt. Over-salting cannot be corrected.
Dredge spices: tandoori masala, coarse cracked coriander seed, black pepper, star anise and garam masala
The flavour of the dredge. Coarse coriander gives visible pieces and the main biltong taste. The masala base sets the profile.
- Swap: Tandoori masala topped up with garam masala and Indian masala when tandoori runs short, as in batch 5 (50 g tandoori + 54 g garam + 50 g Indian masala). Batch 5 did this because it ran short. The record calls batch 5 a set of substitutions, not a clean iteration. Expect a different spice profile.
Cayenne pepper (15.2 g) and mixed dried chilli (33.6 g), toasted and ground to a coarse flake
Heat, red colour and smoke. The mixed chilli is mostly pigment chillies with some chipotle for smoke.
- Swap: Piment d'Espelette in place of cayenne. Espelette is far milder. Batch 5 used 80 g Espelette in place of 15 g cayenne, five times the weight. It is not a like-for-like swap.
MSG, 20 g for 8.7 kg meat, untoasted, in the dredge
Lifts a profile that reads flat when the salt is lower. Works with the inosinate from the anchovy in the Worcestershire.
- Swap: Leave the MSG out, as in revisions 1 to 6. MSG first went in at revision 7. Without it, the profile reads flatter. Do not double it: above roughly 1% of finished weight, glutamate tastes soapy and metallic.
The basic method
8- Trim off the silverskin. Cut the beef along the grain into strips 25 mm thick. Weigh each piece and write the weight down.
- Toast the whole spices and the dried chillies in a dry pan. Cool them. Crush them coarse. Leave the coriander in large pieces.
- Mix the crushed spices with the powders and the salt. Use 15.5 g salt for each 1 kg of raw meat.
- Mix the vinegar and the Worcestershire sauce. Keep the meat cold. Rub the wash into each piece until the surface is tacky.
- Coat each piece in the dredge. Use a lighter coat on thin pieces.
- Put the pieces in the fridge at 3 to 4 °C for 48 hours. Weigh each piece before it goes in.
- Hang the pieces 40 mm apart in a box with a fan that blows the wet air out. Keep the air at 24 to 26 °C. Never let the box go above 34 °C.
- Weigh each piece against its hanging weight. Take a 25 mm piece out when it has lost 46 to 52% of its weight. Trust the scale, not the calendar.
What the full recipe adds
The full recipe sorts pieces into three thickness groups and gives each its own pull target (46 to 52%, 50% and 55%). It starts the thick pieces with a hot phase at 32 to 34 °C for 16 hours to set the bark, then finishes all pieces at 24 to 26 °C. It toasts the seeds in three separate pans because they finish at different times, and it heats the powders at 110 °C for 15 minutes to reduce mould spores. It explains that only about 32 g of dredge sticks to each kg of meat, so salt must be dosed by what adheres, not by what is mixed: batch 7 specified 15.5 g/kg and delivered about 9 g/kg. It gives the measured drying rate of the 25 mm cut (group A in 9 to 11 days) and explains why the 25 mm cut roughly halved the drying time. It explains the chemistry of the red interior and why only nitrite can produce it. These details control salt, mould and colour, which the short pattern cannot.
Batch logs
5The science
1What is actually happening in the dish, and why the techniques work.
The purple centre in an uncured batch is a diagnosis, not a fault, and no amount of adjusting the dry will fix it.
The four states of the pigment
Myoglobin's colour is set by what sits at the sixth coordination site of the heme iron.
- Nothing bound, iron as Fe(II): deoxymyoglobin, purple-red. This is the interior of any piece of meat that oxygen has not reached. It is the purple centre.
- Oxygen bound, Fe(II): oxymyoglobin, bright cherry red. This is bloom on a cut steak. It penetrates only 2 to 3 mm and it is transient.
- Iron oxidised to Fe(III): metmyoglobin, brown. It forms fastest at low oxygen partial pressure, which is exactly the zone just under the surface. This is why stored meat shows a brown band between a red rind and a purple core.
- Nitric oxide bound, Fe(II): nitrosylmyoglobin, stable pink-red.
Drying moves an uncured piece between the first three. It cannot reach the fourth.
Why nitrite is the only route
Nitrite goes to nitrous acid at low pH, and nitrous acid gives nitric oxide, which binds the Fe(II) heme. The NO to heme bond is four to five orders of magnitude tighter than the O2 to heme bond, so the pigment does not need oxygen to stay red and it resists oxidation to met. Because the nitrite diffuses into the meat rather than depending on surface oxygen contact, the colour is uniform through the cross-section instead of being a 3 mm rind. That uniform translucent ruby is what commercial South African biltong shows when it is sliced, and it is not achievable by any drying schedule.
Traditional Boer recipes call for salpeter, which is potassium nitrate. Nitrate is not directly active: bacterial nitrate reductase must reduce it to nitrite first, which is why old recipes need long cure times. A modern nitrite cure salt skips that step and is far easier to dose.
Dosing without moving the sodium
German Nitritpoekelsalz is 0.4 to 0.5% sodium nitrite in sodium chloride, so it is over 99% salt by mass. Substituting it one for one for the salt in the dredge at 13.85 g/kg gives roughly 69 mg/kg ingoing nitrite. That sits inside EU limits for dry-cured meat and above the roughly 50 mg/kg needed for full colour development, at identical sodium. It must be a substitution, not an addition: dosing to 150 mg/kg on nitrite alone would need about 30 g/kg of cure salt, which is more than twice the salt this recipe wants.
Sodium ascorbate at 0.5 g/kg is worth adding with it. It reduces nitrite to nitric oxide far faster, drives the colour all the way to the core during the fridge cure, and suppresses nitrosamine formation.
The vinegar wash helps rather than hinders here, because the low pH is what drives nitrous acid to nitric oxide.
Where this is solid and where it is not
The pigment chemistry is textbook and not in doubt. The specific dosing arithmetic above is calculated rather than measured in this kitchen, and it has not yet been run: nitrite was specified for batch 7 and dropped for lack of supply. The first batch that uses it should record the interior colour deliberately, because that is the whole point of the change.
Two things worth being clear about. Acid accelerates autoxidation of myoglobin to metmyoglobin, so a heavy vinegar wash pushes an uncured interior toward brown rather than red. And myoglobin denatures irreversibly from around 55 degrees, so a heating pad that drives box air past that turns the colour grey no matter what else is done. Keeping the box at or under 34 degrees is a colour control, not only a fat control.
Revisions
7A revision is written when the dish changes, and the version it replaces is kept. Beside each is a sentence describing what the dish became — the part a cook can act on.
First revision cooked and measured rather than planned. Five real changes. (1) Wash collapsed from five vinegars to one: South African brown spirit vinegar, which is the traditional single vinegar, with Worcestershire kept for the glutamate and inosinate the anchovy brings. (2) MSG added at 20 g, roughly a fifth of salt weight, to lift a profile that reads flat once the salt drops. (3) Cut specified at 25 mm along the grain with silverskin removed, rather than whatever the seam gave. This is the largest change and it rewrote the timeline: see the drying rate note. (4) Toast sequence formalised into three pans, and the grind moved to mortar and pestle, which needs about 15% more volume for the same weight. (5) Pull target split by thickness rather than one figure for the batch: 46 to 52% for pieces over 20 mm, 50% for 12 to 20 mm, 55% under 12 mm. Salt raised to 15.5 g/kg to compensate for the wetter pull, since finished salt concentration is green salt divided by finished mass and a 46% pull leaves roughly 17% less salt in the product than the 55% pull of batches 4 and 5. Nitrite was specified and then dropped for lack of supply, so the deep red interior remains unsolved.
Planned 10 kg batch, not yet cooked. Built from batch 4 per-kilogram ratios (the most complete batch without substitution problems), scaled to 10 kg, then batch 5 learnings applied: +30% on the wash and +40% on seasoning volume. Salt deliberately held flat at 13.85 g/kg rather than raised 40% — salt dissolves, so it does not need a volume bump for coverage, and over-salting is the one error here that cannot be corrected.
Ran short on Worcestershire, tandoori masala and cayenne, so this batch is a set of substitutions rather than a clean iteration: 70 g Worcestershire instead of 99.4 g, 50 g tandoori topped up with 54 g garam and 50 g Indian masala, and 80 g piment d'Espelette standing in for cayenne. The two learnings worth keeping are +30% wash and +40% seasoning volume.
Wash finally scaled properly with the meat (~197 g for 7.98 kg). Heat added deliberately for the first time: 10 g cayenne plus ancho, chipotle and guajillo. Tandoori doubled to 107.4 g. The dredge still ran out — 1.6 kg of meat went uncovered.
Added the masala base that defines the profile from here on: 54 g tandoori masala (about a third of total spice weight) plus 3 g garam masala. Coriander raised sharply from 3.53 to 5.92 g/kg. Salt settled at 10% below the batch 1 rate. The wash was left unchanged despite the batch 2 note saying to increase it.
Scaled to 5.98 kg and introduced a wash for the first time — 141 g of mixed vinegars and Worcestershire. Salt cut 20% from batch 1, which was too salty at 15.88 g/kg. Forgot the garam masala entirely and had to add it just before hanging.
Baseline batch. 1.7 kg, no wash at all — just salt, pepper, coriander and star anise straight onto the meat, in two measures because the first was clearly not enough.