Origin

Sourcing Grain: A Visit to Marriages Mill

February 2025

The stone deck oven at Archive Bakery, door open with loaves inside

We drove out to Chelmsford before dawn on a cold Tuesday in February, mostly because we had been buying flour from Marriages for years without ever having seen where it came from. It felt overdue. The mill has been running since 1824, on the same site, and the family that runs it now are descendants of the family that started it — which is a longer, steadier relationship with a piece of land than most bakeries, including ours, can claim with anything.

What struck us first was how much of the operation still depends on human judgement rather than automation. Grain arrives from a network of local farms, and before any of it goes near the mill it is tested and tasted — moisture content, protein levels, and a miller's own sense of whether a batch is right for bread flour or better suited to biscuits. We had assumed, wrongly, that a mill this size would run almost entirely on sensors and software. It does use plenty of both, but the final call on a batch of grain still comes down to someone who has been doing this for decades putting a handful between their fingers.

We spent an hour with one of the millers walking the stone floors, and asked the question we had come to ask: why does flour from a specific mill actually taste different from flour milled somewhere else, given that wheat is wheat? The answer, roughly, is that it is not — the variety of wheat, the soil it grew in, the way it is stored, and the exact temperature and speed of the milling all shift the final flavour. Marriages mills slowly and keeps temperatures low specifically to protect the flavour compounds that heat destroys. That single decision — running the mill slower than it technically could — is the reason their flour behaves the way it does in our dough.

We also learned something that changed how we talk to customers about our country sourdough. The specific wheat variety we use, a heritage strain grown by two farms working with the mill, is not always available. Some years the harvest is smaller, or the protein content is lower than the variety needs to perform well in a long ferment, and Marriages will tell us plainly that a particular batch is not up to their standard for that flour. We have had to adjust our own recipe more than once as a result, mid-season, because the miller was honest about a shortfall rather than sending us a substitute and saying nothing.

Buying from a mill this size costs more than buying from a large industrial supplier, and it means accepting some variation between deliveries that a bigger, blended flour would smooth out. Standing on that mill floor, watching grain that a farmer twenty miles away had grown get turned into the flour that would be in our ovens forty-eight hours later, made the cost easier to justify than any spreadsheet ever has.

Back to the Journal