the Archimedean axiom
/ ar-kih-MEE-dee-an /
Take a thimble and a swimming pool. The Archimedean axiom says something your common sense already trusts: by emptying the thimble into the pool enough times, you can eventually fill it — no length, however large, is beyond the reach of repeated copies of any length, however small. There is no segment so long that a short segment, laid end to end finitely many times, can never overshoot it. It is the axiom that forbids 'infinitely large' and 'infinitely small' lengths from coexisting in the same line.
Stated carefully, it is the first of Hilbert's two continuity axioms. Given any two segments AB and CD with AB the shorter, there is a whole number n such that n copies of AB laid end to end along a ray exceed CD. Equivalently, no segment is 'infinitesimal' relative to another: if you keep adding a fixed small length to itself, the running total grows without bound and eventually passes any target. This is exactly the property of the real numbers that says, for any positive real x and any real y, some integer multiple n times x is bigger than y. It ties the geometric line to the ordinary number line.
It matters because it is what lets you attach real-number lengths to segments at all — measurement, in the everyday sense, depends on it. But it is genuinely an extra assumption, not a logical necessity: there exist consistent 'non-Archimedean' geometries containing infinitesimal segments, where some segment is positive yet smaller than every copy-fraction of a given unit. Those geometries satisfy incidence, order, and congruence but violate this axiom — which is precisely how we know the Archimedean axiom is independent and cannot be proved from the others.
Let AB be one millimetre and CD be one kilometre. The axiom guarantees a whole number n with n millimetres greater than one kilometre — here n = 1,000,001 works. No matter how mismatched the scales, a finite number of small copies always overtakes the large segment.
Enough copies of any small length will overtake any large one — no infinitesimals allowed.
The Archimedean axiom alone does not make the line 'complete' — it rules out infinitesimals but still allows gaps (the rationals are Archimedean yet full of holes). Closing those gaps needs the separate completeness axiom.