A millimetre rule can be read to a millimetre. Past that you are estimating, and everyone estimates differently — which is why two people reading the same rule disagree about the same part. A vernier caliper resolves 0.02 mm, fifty times finer, using no lens, no screw and no electronics. It does it by refusing to make you judge a fraction at all.
The trick is to replace estimating a length with counting a line. That substitution is nearly four hundred years old, it is entirely geometric, and once you have seen it once the formula everyone memorises stops being something to memorise.
Cut fifty divisions to fit forty-nine millimetres
Take a second, sliding scale and engrave 50 divisions on it — but grind them to span only 49 mm of the main scale, not 50. Each vernier division is therefore 49 ÷ 50 = 0.98 mm, which is exactly 0.02 mm short of a main-scale millimetre.
That shortfall is the whole instrument. It is also, by definition, the least count. The figure below uses the coarse 10-division scale rather than the fine one, because a 0.02 mm shortfall is about one pixel wide at any honest scale, while a 0.1 mm one is a tenth of a division and impossible to miss. The arithmetic is the same either way.
Only one line can ever be flush
Slide the vernier right by some fraction of a millimetre. Its zero no longer sits on a main-scale line — that is the fraction you cannot read. But look along the vernier: line 1 is one least count behind where a main line sits, line 2 is two behind, line 3 three, and so on. Push the whole scale forward by k least counts and line k — and only line k — closes its gap exactly and comes flush.
So the fraction is no longer a length to judge. It is a line number to count, and the eye is far better at spotting two lines that meet than at guessing a proportion. Its neighbours miss by one least count in opposite directions, which is the visual signature to look for.
Which is where the formula comes from
The reading is the whole millimetre the vernier zero has passed, plus the flush line's number times the least count:
total = main scale + (vernier division × least count)
With the jaws at 24.34 mm on a 0.02 mm caliper, that is 24 + (17 × 0.02) = 24.34 mm. Nothing was estimated: 24 was read off a millimetre scale, and 17 was counted.
There is a tidy identity hiding in that sum, and it is worth knowing because it tells you where to look. The flush line lands on main-scale division 24 + 17 = 41 mm — the main-scale reading plus the division number, in millimetres. So on a 50-division caliper, the coincidence is never more than 50 mm to the right of the vernier zero, and its distance from the zero counts the hundredths directly.
The least count is a design choice, not a constant
Nothing above depended on the numbers 50 and 49. Grind N divisions to span N − 1 millimetres and the same argument gives a least count of 1/N mm. Written generally, with MSD standing for one main-scale division:
least count = MSD × (1 − span ÷ divisions)
| Vernier | Divisions | Span | One division | Least count |
|---|---|---|---|---|
| 50-division metric | 50 | 49 mm | 0.98 mm | 0.02 mm |
| 20-division metric | 20 | 19 mm | 0.95 mm | 0.05 mm |
| 10-division metric | 10 | 9 mm | 0.9 mm | 0.1 mm |
| 25-division inch | 25 | 0.6 in | 0.024 in | 0.001 in |
The inch caliper is the same construction wearing different clothes. Its main scale is divided into fortieths of an inch — 0.025 in a division — and 25 vernier divisions are ground to span 24 of them. That gives 0.025 × (1 − 24/25) = 0.001 in, which is a thousandth of an inch. No new idea, just a different MSD — though carrying a reading between the two systems afterwards is a job for the length converter rather than for the scale.
Why nobody sells a 200-division vernier
If 1/N mm is available for any N, why stop at fifty? Not because of the eye. Detecting whether two abutting lines are aligned is a thing human vision is startlingly good at — thresholds of two to five arcseconds, well under the width of a single cone in the retina. Vision science calls the ability vernier acuity, and the name is not a coincidence. At normal reading distance that is a misalignment of roughly six thousandths of a millimetre.
What stops it is geometry. A vernier of N divisions is physically N − 1 main-scale divisions long, because that is the definition. So the scale itself grows as the least count shrinks — and it has to sit on a beam that is only 150 mm long in the first place.
There are two more limits behind that one. The instrument's own errors — jaw parallelism, beam straightness, the force your thumb applies — are already worth several hundredths of a millimetre, so a scale claiming five thousandths would be reporting precision the hardware does not have. And searching two hundred nearly identical lines for the flush one is slow and easy to get wrong, which is the opposite of what a shop-floor instrument is for.
So finer measurement changed mechanism instead. A micrometer converts a 0.5 mm screw pitch into a full turn of a thimble, spreading fifty graduations around a circle rather than along a beam — 0.01 mm read like a clock face, in a device with no length problem at all. The vernier kept the range; the screw took the resolution.
Ninety years from the idea to the instrument
The ancestor is the nonius, published in 1542 by the Portuguese mathematician Pedro Nunes. To subdivide a quadrant he engraved a family of concentric arcs, the outer one cut into 90 parts, the next 89, the next 88, and so on down — so a pointer that fell between graduations on one arc would fall on a graduation of another. The idea is right and the instrument is a nightmare: forty-odd separate scales, and interpolation between them anyway.
In 1631 the French mathematician Pierre Vernier (1580–1637) replaced the whole family with a single sliding scale of the type above, in a book on a new mathematical quadrant. It took until the middle of the eighteenth century to come into common use, and until the end of that century for Jérôme Lalande to attach Vernier's name to it. The Portuguese word survives too: in German and Dutch a vernier scale is still a Nonius.
Watch it happen
All of this is easier to believe with a scale in your hand. The vernier caliper simulator draws the instrument at all four least counts in this article, marks the flush pair for you until you switch the hint off, and writes the sum out as you drag. Set it to 0.1 mm first, where the shortfall is a tenth of a division and you can watch the coincidence walk along exactly as in the second figure, then drop to 0.02 mm once the pattern is obvious.