the Dandelin spheres
/ dahn-duh-LAN /
Here is a genuine 'aha' of geometry. The Greeks knew conics two ways — as slices of a cone and as curves defined by foci — but it was not obvious these were the same curves. In 1822 Germinal Dandelin found a breathtaking proof using two spheres tucked inside the cone, and it is one of the most beautiful arguments in all of elementary geometry.
Picture slicing a cone with a plane to make, say, an ellipse. Now slip a sphere into the cone above the cutting plane, inflating it until it just touches the plane at one point and snugly touches the cone all around a circle; then slip a second sphere in below, doing the same. These are the Dandelin spheres. The claim: the two points where the spheres kiss the plane are exactly the foci of the ellipse. The proof rests on one elementary fact — the two tangent lines from an external point to a sphere have equal length. Take any point P on the conic and draw the line from the cone's apex through P down the surface; this line touches the upper sphere where its tangent circle sits and the lower sphere likewise. The distance from P to the upper focus equals P's tangent length to the upper sphere, which equals the distance along the cone's surface from P up to the upper tangent circle; similarly for the lower. Adding them gives the distance between the two tangent circles along the cone — a constant that does not depend on which P you chose. So |PF_1| + |PF_2| is constant: the slice satisfies the focal-distance definition.
This single idea ties all three definitions together: the same construction works for the parabola (one sphere) and the hyperbola (one sphere in each nappe), and even pins the directrix at the plane where the sphere's tangent circle lives. It is the proof that 'slice of a cone' and 'curve with foci' name the same objects — no coincidence, but a theorem.
For an ellipse, the two inscribed spheres touch the slicing plane at points F_1 and F_2. For any P on the ellipse, |PF_1| equals the slant distance up the cone from P to the upper sphere's contact circle, and |PF_2| equals the slant distance down to the lower circle. Their sum is the fixed slant gap between the two contact circles — the same for every P, which is exactly 2a.
Equal tangent lengths turn the sum |PF_1| + |PF_2| into a constant — the focal definition, proved.
The whole proof leans on just one lemma: all tangent segments from a fixed external point to a sphere have equal length. It is not magic — it is that single elementary fact applied twice, once per sphere.