the angle of parallelism
In Euclid's plane, distance and angle are unlinked: a given length tells you nothing absolute about any angle. The angle of parallelism is the startling fact that in HYPERBOLIC geometry they are welded together. Drop a perpendicular of length p from a point P to a line L. Now swing a ray from P down toward L. Rays steep enough hit L; rays too shallow miss it entirely; and there is one exact borderline ray that just barely fails to meet L — the limiting parallel. The angle that this borderline ray makes with the perpendicular is the angle of parallelism, written Pi(p). It depends ONLY on the distance p.
Concretely, Pi(p) is always an acute angle (strictly less than 90 degrees), and it shrinks as p grows. When P is very close to L (p small), Pi(p) is just under 90 degrees and the picture looks nearly Euclidean — the limiting parallel comes off almost square to the perpendicular. As P moves far from L (p large), Pi(p) drops toward 0 degrees: the borderline ray leans steeply along the perpendicular, and the fan of non-intersecting directions opens wide. Lobachevsky's exact formula is tan(Pi(p) / 2) = e^(-p/k), where k is the curvature constant of the plane; only at the unreachable Euclidean limit (k -> infinity) does Pi(p) stay locked at 90 degrees for every p.
Why it matters: this single function is the quantitative heart of hyperbolic geometry. Because Pi(p) is determined by p alone, length acquires an ABSOLUTE meaning — you can name a specific length purely by the angle it produces, with no arbitrary ruler. That is impossible in Euclid, where you may freely rescale everything, and it is the deep reason similar-but-unequal triangles cannot exist in the hyperbolic plane.
Set the curvature constant k = 1. At distance p = 1 from the line, tan(Pi/2) = e^(-1) = 0.368, so Pi(1) = 2 arctan(0.368) is about 40.4 degrees. At p = 2, e^(-2) = 0.135 gives Pi(2) about 15.4 degrees. Doubling the distance more than halves the parallelism angle — the further you stand from a line, the more steeply its parallels lean.
The borderline 'just-misses' ray makes the angle Pi(p) with the perpendicular; it depends on distance alone and shrinks toward 0 as you move away.
In Euclid the angle of parallelism is a constant 90 degrees for every distance, which is why no such relation is ever taught there. Its variation with distance is unique to hyperbolic geometry and has no Euclidean counterpart.