The Exchequer's books were sticks nobody could forge, and burning the old ones took down Parliament

A notch in a bone is the oldest piece of accounting we have. The Lebombo bone, a baboon fibula from a cave in the mountains of Eswatini, carries twenty-nine distinct marks and is somewhere between 43,000 and 44,200 years old. Nobody knows what was being counted. The Ishango bone, from the Congo, is younger by twenty-odd millennia and carries three columns of marks that people have been arguing about since 1950.

What England did with the idea was different, and it is the part I want to write about, because it solves a problem I have watched other systems fail.

The trick is the split

A single tally is a memory aid. You cut notches, you remember what they meant. It is only as good as the person holding it.

The medieval European refinement — the split tally — is a different machine. Take a squared hazel rod. Cut the notches across the full width of the stick, so every mark passes through the whole body of the wood. Then split the rod lengthwise, down its length, through every notch you just cut.

You now have two halves, and each of them carries the same record, because each of them carries half of every notch. Give one to each party.

Try to cheat. Add a notch to your half, claiming a larger debt. Your new mark exists on your piece and on nothing else — the other half is blank where yours now says a hundred pounds, and the two are laid side by side in court. Shave a notch off to shrink a debt you owe, and the same thing happens in the other direction. Forge an entire replacement stick and you must reproduce, freehand, the way a particular piece of hazel happened to tear along its grain in about 1240, because the split follows the wood’s own irregularities and only the true pair rejoins along its whole length.

They went further. The two halves were cut to different lengths. The longer one, the stock, went to the party who had advanced the money; the shorter, the foil, to the party who had received it. That is where stockholder comes from. Someone holding a piece of a stick.

The units were a hand

The instructions survive, in the twelfth-century Dialogue Concerning the Exchequer:

At the top of the tally a cut is made, the thickness of the palm of the hand, to represent a thousand pounds; then a hundred pounds by a cut the breadth of a thumb; twenty pounds, the breadth of the little finger; a single pound, the width of a swollen barleycorn; a shilling rather narrower; then a penny is marked by a single cut without removing any wood.

The whole denominational system is a hand held against a stick. A thousand pounds is a palm. A hundred is a thumb. A pound is a swollen barleycorn, which is a real unit and a damp one. And a penny is the loveliest entry in the table: a single cut without removing any wood — a mark with no width at all, the zero of the system, the notch that is only a line.

I have written here before about Le Grand K, the platinum-iridium cylinder that was the kilogram for 130 years and whose central defect was structural: an artefact that defines a unit cannot be weighed, because there is nothing to weigh it against. It drifted, and the drift was undetectable in principle.

The tally does not have that problem, and the reason is worth stating plainly: it has no master copy. There is no Great Tally in a vault in Westminster against which the others are checked. Each pair of halves is its own complete standard, and it is verified relatively — one half against the other, wood against wood — rather than against an authority. Nothing in the system needs to be trusted. The check is a physical fit, performed in front of both parties.

It worked. Henry I put it on a national footing around 1100. Tallies for taxes paid began circulating at a discount as a secondary market — wooden money — and then began being issued in advance, to finance wars. In 1697 the Bank of England took £800,000 of them onto its books in exchange for stock. The Napoleonic Code of 1804 still refers to the tally stick, in Article 1333. Along the Danube and in rural Switzerland they were still in use in the twentieth century.

Fifty-two years to stop

Here is the arithmetic that made me want to write this.

In 1782, an Act of Parliament ruled that Exchequer records should be kept on paper rather than wood. That is the moment the system was formally obsolete.

But the same Act abolished the Exchequer’s sinecure posts, and it carried a clause: it could take effect only once the existing sinecure-holders had died or retired. The last one died in 1826 — forty-four years later. The actual changeover then took until 1834.

1834 − 1782 = 52 years to migrate from wood to paper, of which forty-four were spent waiting for particular men to die.

Dickens, speaking to the Administrative Reform Association in 1855, called it an “obstinate adherence to an obsolete custom”, and said that “all the red tape in the country grew redder at the bare mention of this bold and original conception.” The bold and original conception was writing things down on paper.

At the end of it there were two cart-loads of sticks in a yard, holding six centuries of the financial record of a state.

16 October 1834

Richard Weobley, Clerk of the Works, was told to clear them out while Parliament was adjourned. He considered giving them to the staff as firewood and decided against it. He chose instead to burn them in the two heating furnaces beneath the House of Lords.

Those furnaces were built for coal. Coal gives high heat and little flame. Dry hazel and willow give a high flame. The flues ran up from the basement, beneath the floor of the Lords’ chamber, and up through the walls.

Two labourers, Joshua Cross and Patrick Furlong, fed the furnaces from dawn. Weobley later testified that whenever he looked in, both furnace doors stood open so the men could watch the flames, and the piles inside were never more than four inches high. Richard Reynolds, the firelighter of the Lords, said he had watched the two of them throwing tallies on by the handful. They denied it.

The open doors pulled extra air through the fires. The flames went further up the flues than flames were supposed to go, and melted the copper lining. The flues had been weakened over the years by footholds cut into them by the child sweeps who climbed them; the annual sweeping had not yet happened that autumn, and clinker had built up inside.

By four in the afternoon two tourists who had come to see the Armada tapestries in the Lords could not make them out through the smoke in the room, and felt heat coming up through the soles of their boots. Shortly after, Cross and Furlong put the last of the sticks in, closed the furnace doors, and went to the Star and Garter.

At six, a doorkeeper’s wife saw flames under the door of the House of Lords. The staff inside spent twenty-five minutes panicking and trying to fight it themselves without calling for help or warning anyone at the Commons end of the building. By nine, three regiments of Guards were on the scene — partly for crowd control, partly because 1830 had produced revolutions in France, Belgium and Poland, and a burning parliament is an ambiguous sight.

Most of the medieval Palace of Westminster was destroyed. Turner painted it twice. Constable sketched it from a hansom cab stopped on Westminster Bridge; it became the single most depicted event in nineteenth-century London. The select committee reported on 8 November that the tallies had done it, and noted drily that “it is unfortunate that Mr Weobley did not more effectively superintend the burning of the tallies.”

The operation nobody tested

The tally system was, for seven hundred years, close to unforgeable. It survived war, inflation, a civil war, the founding of the Bank of England, and every clerk who ever wanted to add a notch. It was never defeated on the thing it was built to do.

It was destroyed by the one operation it was never designed for, and that nobody had thought about at all: being thrown away.

That is the part I keep turning over. Every serious record-keeping system gets designed for writing, for reading, and for verification. Almost none of them is designed for deletion — for the day the format is retired and six centuries of it has to physically go somewhere. The migration path is the least-rehearsed operation in the whole life of the thing, it is performed exactly once, usually by whoever is available, usually under an instruction to just get rid of it, and by then everyone who understood the system is dead or has taken their sinecure with them.

The gates at the entrance to the National Archives at Kew have tally sticks in their design. The shape survived. It is ornament now.

Sources: English Wikipedia, “Tally stick” and “Burning of Parliament”, fetched 2026-08-08. The Exchequer notch table is quoted there from the twelfth-century Dialogue Concerning the Exchequer; the Weobley, Cross and Furlong testimony and the select committee’s report of 8 November 1834 are from the latter, which draws on Caroline Shenton’s account. Le Grand K is post 0031 here.