a Lambert quadrilateral
/ LAM-bairt /
Johann Heinrich Lambert, around 1766, pursued the same strategy as Saccheri but with a tidier figure. A Lambert quadrilateral is a four-sided figure with THREE right angles. The single remaining 'fourth' angle is then the whole story: in Euclid it must also be 90 degrees (making a rectangle), but if the geometry is non-Euclidean that fourth angle is forced to be either obtuse or acute. It is exactly half of a Saccheri quadrilateral, cut along its axis of symmetry — which is why the two figures carry the same information.
Walk the logic. With three right angles fixed, the fourth angle alone decides everything. Fourth angle right => Euclidean, the figure is a rectangle, and the two sides adjacent to the fourth angle equal the two opposite sides. Fourth angle obtuse => spherical/elliptic geometry. Fourth angle acute => hyperbolic geometry, and here Lambert noticed something prophetic: the AMOUNT by which the angle falls short, the defect, is proportional to the AREA of the quadrilateral, and the geometry behaves as if drawn on a sphere of imaginary radius. He even saw that this would force an absolute unit of length — a fact with no Euclidean analogue.
Like Saccheri, Lambert hoped to find a contradiction in the acute case and never did, though he was more honest about his failure to do so. His quadrilateral matters because it isolates the parallel question into a single angle and ties the non-Euclidean defect directly to area — the seed of the theorem that a hyperbolic triangle's area equals its angle defect. Both quadrilaterals are tools of NEUTRAL geometry: every theorem about them up to the three-cases split holds without assuming any parallel postulate at all.
Take a quadrilateral PQRS with right angles at P, Q, and R. In Euclid the angle at S is forced to 90 degrees and PQRS is a rectangle. In the hyperbolic plane the angle at S comes out acute, say 84 degrees; the 6-degree shortfall is proportional to the area enclosed, and the sides flanking S are longer than their opposites.
Three right angles fixed; the fourth angle — right, obtuse, or acute — names the geometry, and its defect tracks area.
A Lambert quadrilateral is simply half a Saccheri quadrilateral, so they encode identical facts; neither is the 'real' figure. The acute fourth angle is not a flaw — it is hyperbolic geometry behaving exactly as it should.