A steelyard is a plain wooden or brass beam with a hook at one end, a loop of cord somewhere near that end, and a small weight that slides along the long side. Hang a sack of rice on the hook, hold the loop, slide the weight until the beam hangs level, and read the number under it. No springs, no pile of matching weights, no pair of equal pans. The strange part is that the sliding weight is small — often a few hundred grams — while the load may be tens of kilograms. It is not a trick of the material. It is what the position of the pivot buys.
Take the cord loop as the fulcrum. The load hangs at a short distance on one side, the weight at a variable distance on the other. Balance is a matter of moments: load times its short arm equals weight times its long arm. If the hook sits four centimetres from the loop and the weight can travel eighty centimetres along the beam, the long arm is twenty times the short one, so a half-kilogram weight balances ten kilograms. Nothing has been made lighter; the force has simply been given a longer lever to work through.
This is why a steelyard is not really a weighing machine in the sense a spring scale is. A spring measures force by how far it deforms. A steelyard compares two moments and asks only whether the beam is level — a yes-or-no question the eye is extremely good at answering. The reading comes from the position at which the answer switches from one side to the other. Level is the null point; the scale is a device for finding a null point quickly.
Why the marks are evenly spaced
The elegant part follows from the same equation. For a fixed weight and a fixed short arm, the distance the weight must travel is directly proportional to the load. Double the load and you double the distance. So the marks along the beam — traditionally small inlaid stars or dots of brass — are equally spaced. A maker does not need a table of squares or a gradual compression of the graduations. One known weight is hung, the balance point is marked, another equal load is added, and the next mark falls the same distance further along.
In practice the zero is not at the loop. The beam itself has weight, the hook and its chain have weight, and the beam is often tapered, so even with nothing hanging on it the whole assembly has a moment about the cord. The empty balance point is found first and becomes the first star; the scale is counted from there, not from the fulcrum. This is also why a worn or repaired steelyard cannot simply be re-inked from a ruler: changing the hook, adding a new chain or planing the beam shifts the zero and every mark with it.
Makers calibrate rather than calculate. A standard weight is hung, the sliding weight is moved until the beam is level, and a point is inlaid; the process repeats down the length of the beam. The taper of the wooden arm is part of this — it keeps the beam lighter at the far end so its own moment does not grow too quickly, and it gives the hand a clear sense of which end is which. Fine marks at the light end and coarser ones at the heavy end are possible, though the underlying spacing stays linear.
Some beams carry two cord loops a few centimetres apart. Changing loops changes the fulcrum and therefore the short arm, which changes the whole range: the loop nearer the hook gives a shorter load arm and a larger capacity with coarse, crowded readings, while the farther loop gives a smaller capacity with finer ones. It is a two-range instrument made with a second piece of string. The same trick appears in any trade where one tool has to weigh both a bundle and a handful.
Why a small weight is the right choice
A heavier sliding weight would shorten the beam needed for the same maximum load, which sounds like an improvement. It is not. A heavy weight compresses the graduations: each extra kilogram moves it a smaller distance, so the same unsteady hand or the same worn mark represents a larger error in the reading. A light weight spreads the scale out and makes the beam readable to a fraction of a unit. The cost is length — the beam must be long enough to let that light weight travel all the way out — and a long beam brings its own weight and its own bend. Range, resolution and length are one decision, not three.
The limit shows up at the far end. A very long beam sags, and a sagging beam changes the geometry it was calibrated for; it also becomes awkward to carry and easy to knock out of true. Makers stop well before that point, which is why a market steelyard has a definite maximum and why larger loads move to a different tool altogether. The small weight is chosen last, after the length and the range have been settled.
Other objects here turn a hard quantity into an easy position. An abacus turns a number into a pattern of beads against a beam, read at a glance and undone with one push. A loom stores a whole repeat in its shafts, so the weaver follows a sequence instead of remembering a design. A jianzi puts its mass at the bottom so one attitude is overwhelmingly likely. Each one moves the difficulty out of the user's head and into the layout of the thing.
The steelyard's version is a straight line of evenly spaced stars. Once that line has been set against a known weight, no one using the beam has to understand moments at all: hang, slide, level, read. The physics is done once at the bench where the marks are inlaid, and every weighing afterwards is just finding the dot where the beam stops tilting.