Market day rotation

Why do neighbouring market towns hold their markets on different weekdays instead of all choosing the busiest day?

Look at an old county map with its market days written beside the towns and the days are all different — Monday here, Tuesday four miles off, Thursday over the hill. The same arrangement turns up in West Africa, rural China and medieval Europe, among people who never compared notes.

If some day is genuinely best for trade, everyone should want it. Instead neighbours take turns, as though by agreement, and there is rarely a record of any agreement. So either this is decorative custom, or the towns are solving a problem we have not yet named.

There are two puzzles here, not one. Why hold a market on some days rather than keeping a shop open on all of them? And, given that, why do neighbours pick different ones?

Take the first. A seller needs a minimum catchment to live on — the threshold. Spread a thin rural population across a landscape and no single village reaches it: a village of a few hundred cannot keep a cloth merchant. But a merchant who moves adds villages together. Six places, each supplying a sixth of a living, make one living if he visits each in turn. Periodicity is not quaintness; it is the device that lets a trade exist where no fixed shop could.

A second reason sits on the buyers' side and concerns coordination rather than income. A market works only if enough buyers and sellers are present at the same moment — sellers with no customers fail as surely as the reverse. A fixed, known day solves that: nobody has to guess.

Now the second puzzle. Two pressures push the same way. The trader has one body, so he can serve six villages only if their days are distinct; if two clash he abandons one, and that village loses its market rather than sharing it. And a buyer within reach of two markets on one day must choose. Both are then thinner, and thinness feeds on itself — sellers go where the buyers are and buyers where the sellers are, so the weaker of two clashing markets does not merely lose custom, it can die.

Both pressures depend on distance, and only on distance. Two markets far enough apart that no trader could reach both and no buyer would weigh both may share a day happily. The rule is therefore local: markets within a day's travel must differ, markets beyond it need not.

That is a familiar structure. Draw a dot for each market, join any two within a day's travel, then hand out days as though they were colours with no two joined dots taking the same one. The number of colours needed is set by how many markets crowd inside one trader's range — which is why market weeks are not all seven days. West African systems ran on four-, five- and eight-day weeks; rural China used a ten-day cycle with schedules such as 1-4-7 for one market and 2-5-8 for its neighbour, as Skinner's study of Chinese marketing describes it in my recollection.

How far is a day's travel? Derive it. A buyer must walk there, trade, and walk home before dark: at roughly three miles an hour, with four or five hours spent walking, that is six or seven miles each way. Medieval English practice put it at six and two-thirds miles — a third of a twenty-mile day, attributed to Bracton, as I recall it — which is the same arithmetic rather than a coincidence.

This sits beside the argument about how far apart towns space themselves. There, thresholds and ranges are satisfied in space: a service appears wherever enough customers live inside its range, and settlements sort into levels. Here the threshold is met at no single point, so it is satisfied in time instead, by pooling sub-threshold places into one itinerant living. That is the fixed point of difference: spacing is the spatial solution to a threshold, periodicity the temporal one, and the same two quantities generate both.

Nor is the rotation purely emergent, against any reading of it as a natural pattern of the landscape. In England markets existed by royal charter, and a charter could be challenged in court on the ground that a new market injured an existing one nearby — the six-and-two-thirds-mile rule was a legal instrument, not a description. An institution was enforcing what the walking distance implied.

Three qualifications. Some clashes are deliberate: a large centre with real pull may take the best day to dominate a smaller neighbour. Religious observance shortens the usable palette, so the constraint sometimes binds harder than range alone suggests. And in some systems buyers circulate rather than sellers, which changes who bears the cost of the timetable but not its logic.

What would refute the account? It makes a cheap prediction: map a dense region's market days and the assignment should be strongly anti-correlated with distance — near neighbours almost never sharing, distant pairs sharing freely. If a well-recorded region showed adjacent markets clustering on the same weekday, the argument is finished. A second test runs through time: if periodicity exists because thresholds are unmet locally, rising density and cheaper transport should dissolve it. English towns that kept the charter and lost the market are the residue of exactly that.

• How periodic markets nest into a hierarchy, with larger centres holding more frequent or higher-order markets.
• Why annual fairs follow a different logic, timed to harvests and livestock rather than to walking distance.

Every term the collection defines is gathered in the glossary.