Foeder Pale Sour – Lark Barrel 2026

About This Beer

A pale, oak-aged sour for the long game, in an ex-Lark Distillery barrel from Dave Bell at Garage Project’s Wild Workshop. Neutral oak with very little spirit character left in it, so the wood is a home for the culture rather than a flavour of its own. Aiming for a clean, bright, Rua-style foeder pale sour: soft acidity, gentle funk, a continental-pilsner backbone that lets the culture lead.

Inoculated once, at the first fill, with a 375 mL bottle of live culture pulled straight from Garage Project’s 8,000 L Rua foeder.

This page is the whole barrel. Five brew days across two months produced eight ~23 litre cubes of one identical wort. Seven went into the barrel and the eighth was diverted onto Sixth Magic before the barrel arrived, rather than sit indefinitely in a sealed cube. The ingredient figures here are the totals for the seven, not any one brew day.

The same wort also runs on into Second Sight, brewed on the Sixth Magic cake in August.

The wort and the method

Every cube is the same 5.5 kg build: continental pilsner base, Weyermann wheat for body and head, rolled oats for a mouthfeel that survives long attenuation. Brewed no-chill, run hot into sealed cubes, held until its round came up.

Two substitutions crept in. A stock shortage mid-run meant two cubes ran part standard Weyermann Pilsner in place of Barke, and one single cube used it as the whole base. The bittering hop switched from Pacific Jade to Simcoe on the later cubes. At 60 minutes and around 15 IBU, bittering only, both sit happily in the same barrel.

Ferment first, then transfer. Running primary inside the barrel would leave a thick yeast bed sitting in the wood for weeks and risk autolysis, so every round fermented out in the 60 L all-rounder and only mostly-clear beer moved across. The barrel gets a CO2 purge before each fill and again on the headspace afterwards. Each round goes over just shy of finished, so it keeps throwing a little CO2 inside the barrel and holds the whole thing positive while the next rounds are brewed.

The fill

Four rounds, all on Lallemand BRY-97, a clean cold-tolerant strain that secures the beer quickly in a winter garage. Rounds 1 to 3 ran on a single pitch, each one going onto the last one’s cake.

Round 1, transferred 16 June. Two cubes, about 44 L. The 375 mL of live Rua went in with this fill and this fill only.

Round 2, transferred 19 June. Two cubes on the round-one cake. Fastest ferment on record here, down to 1.045 by 4:30pm on pitch day, and it overflowed to a blow-off.

Round 3, transferred 26 June. The held batch-1 cube plus the first single batch, on the same cake a third time, run cooler and finished more fully. Bung off and CO2 purge during setup, lines purged, gas-pushed from the all-rounder at 4 to 5 psi with the sanitised line submerged below the surface. Rounds 1 to 3 combined at 1.057 down to 1.010, about 6.2%, on one RAPT Pill.

Round 3 was meant to leave 10 to 15 L over to keg as top-up reserve. Instead the barrel took the entire round, roughly 125 L of clean beer, and was still about 10 L short of full. No reserve banked, and a fourth round needed.

Round 4, brewed 13 July, transferred 22 July. A single cube back on the original Barke base with a fresh BRY-97 pitch, chilled into a FermZilla rather than the all-rounder. 1.061 down to 1.014 over eight days, about 6.2%. Three or four points higher than rounds 1 to 3, which is what a light fresh pitch on a higher OG does, and immaterial in a barrel where the Rua bugs set terminal gravity anyway.

It took only 7 to 7.5 L to brim the barrel, so about 132 L of beer went in across the four rounds against a nominal 115.

That is beer poured in, though, not internal capacity, and the two are not the same on a first fill. A quarter cask is nominally 125 L, and this barrel came in dry: the staves wept on the lower half for weeks afterwards, and the white mould that had to be cleared off the outside was feeding on beer that had seeped right through the wood. Dry oak takes up several litres before it swells shut. So the likely reading is a standard 125 L quarter cask plus roughly 7 L absorbed into and through the staves, rather than an oversized barrel.

Worth knowing for the next fill, because that absorption is a one-off. The wood is saturated and swollen shut now, so the barrel should want closer to its nominal 125 L next time, not 132.

The balance of round 4, about 13.5 L, went into a purged keg as the airlocked angel’s-share reserve that round 3 failed to leave.

Now aging

Fill phase over. A little egg-like sulphur came off the earlier fills, benign fermentation sulphur that gases off. The pin in the bung has been swapped for a bubbler for a clearer view of activity and a little more protection against ingress. A low fan across the staves cleared an earlier weep-damp and white mould on the lower staves; the exterior is dry to touch and the staves have swollen shut.

No pellicle yet, and none expected this early. A 375 mL culture into 132 L is roughly a 1:350 pitch, in a garage sitting near 10°C, so film development is slow and may stay patchy or absent while the beer sours perfectly well beneath it. Brim-full at 6.2% with a falling pH is protective; the real risk is headspace oxygen turning acetic, which is what the brimming and the reserve top-ups defend against.

Twelve to eighteen months, taste-driven. First sample around late October 2026 for a baseline pH and taste, then every three to six months, topping from the reserve keg each time. Target finished pH 3.0 to 3.3. It comes out when acid, funk and oak sit in balance, which is a decision for the palate rather than the calendar.