An abacus looks like a tray of beads, but it is really a machine for holding place value. Each vertical rod is one decimal position — units, tens, hundreds — and a horizontal beam divides every rod into a short upper section and a longer lower one. Beads only count when they are pushed against that beam. Beads sitting away from it are simply parked. That single rule turns a pile of identical wooden discs into a readable number without a single written mark.

The values are unequal. Each of the five lower beads is worth one. Each of the two upper beads is worth five. So a rod can show any value from zero to fifteen: five ones, plus two fives. The upper beads exist because five is the half of ten that a hand can grasp without counting — four lower beads pushed up and one upper bead pushed down are the same value, but the upper bead is one movement instead of four. Grouping is what makes the frame fast.

Why two fives and five ones, when decimal arithmetic never needs a digit above nine? Because the frame was also used for a different arithmetic. Under the old weight system one pound was divided into sixteen liang, so converting pounds and liang, or pricing goods quoted by the liang, meant working in base sixteen on the same rods. A single digit in base sixteen runs from zero to fifteen. Two upper beads and five lower beads give exactly that range. The same instrument could therefore do ordinary decimal bookkeeping and the sixteen-part weight arithmetic without any change of tool.

What the beam actually does

The beam is a read line. An abacus is not read by counting all the beads on a rod; it is read by looking at which beads touch the beam. This is a small but decisive piece of design: the eye takes in a pattern rather than a quantity, and a mistake is visible as a gap in the wrong place. On paper, an intermediate result has to be written down, and correcting it means crossing it out. On the frame, an intermediate result is a physical arrangement that can be undone by pushing one bead back.

Carrying is a fixed gesture rather than a calculation. When a rod passes nine, the operator clears it and adds one to the rod on the left. Subtraction borrows in the reverse direction. Skilled users do not recite arithmetic while doing this; their fingers run learned sequences, much as a musician runs a scale. Speed comes from the fact that the rules have been built into the layout — the rods are ordered, the beam marks what counts, and the bead values never change. What is left for the person is repetition.

Zero costs nothing and needs no symbol. An empty rod, all beads pushed away from the beam, is simply a zero in that position, and it holds the place for every position to its left. This is why the number of rods matters more than the number of beads: a longer frame can hold a larger number, but it cannot hold a more precise one. Precision is not a feature a counting frame has; it works in whole units, and the operator decides in advance what one unit means — a coin, a liang, a tenth of a measure.

Why beads rather than marks

The frame is also a memory that cannot be smudged. A shopkeeper weighing goods, calling out a price and being interrupted can leave the running total sitting on the rods and come back to it. That is a practical advantage no written column of figures has, and it is the reason the abacus stayed in use long after paper and pencil were cheap. It is reversible, portable and readable from a glance at an angle, and it needs neither ink nor light beyond what the shop already has.

The craft is in the frame, not the beads. Rods must be parallel and close enough that the beam line stays straight; beads must slide freely but not so freely that a knock changes the number. Hardwood, a slightly oval bead hole and a beam with a small clearance are all choices about friction. A frame that is too tight tires the hand over a day of accounts; one that is too loose loses a figure when someone sets a bowl down beside it. Makers judge this by running a bead the length of a rod and listening.

Other crafts on this site work the same way, putting a rule into the object so the person does not have to carry it in their head. A loom stores the pattern in its shafts and treadles; a woodblock stores a whole page as one raised surface. The abacus stores place value. In each case the skilled action is not deciding what to do but doing a short sequence correctly, and the object makes the wrong sequence physically awkward.

A steelyard balance is the closest relative among measuring tools: it converts weight into a position along a rod, so the answer is read as a distance rather than computed. The abacus converts number into a position along a row of rods, so the answer is read as a pattern. Both let a merchant settle a question without writing anything down.

If the frame were used only for decimal arithmetic, one upper bead would be enough — five ones plus one five already covers zero to nine. The second upper bead and the fifth lower bead are the surplus that makes fifteen possible. That surplus is the visible trace of a different counting system sharing the same bench, preserved in wood long after the reason for it became less common. The shape outlasted the arithmetic that explains it.