The vector A=r˙×h−GMr^ is constant in time for motion under an inverse-square force r¨=−r3GMr, has magnitude ∣A∣=GMe, and points from the focus toward periapsis.
Why is it true?
Energy and angular momentum alone fix the size and shape of a Kepler orbit but not its orientation within the plane; the Laplace–Runge–Lenz vector is the extra conserved quantity that pins down where periapsis sits, and its very existence is special to the 1/r2 force — most central forces do not have such a vector, which is exactly why most central-force orbits (as in the restricted three-body problem below) do not stay on a fixed closed curve.
Proof sketch
Differentiate: dtdA=r¨×h+r˙×h˙−GMdtdr^. Since h is constant (Theorem 1), the middle term vanishes. Using r¨=−r2GMr^ and h=r×r˙, the vector triple product identity r^×(r×r˙)=r(r^⋅r˙)−r˙(r^⋅r) gives r¨×h=−r2GM[rr˙r^−rr˙]=GM(rr˙−rr˙r^), where r˙=r^⋅r˙ is the scalar rate of change of the distance r (used r^⋅r=r).
On the other hand, differentiating r^=r/r directly gives dtdr^=rr˙−r2rr˙=rr˙−rr˙r^ — exactly the same bracketed expression found for r¨×h/GM above. So r¨×h=GMdtdr^ identically.
Therefore dtdA=GMdtdr^−GMdtdr^=0, proving A is constant. Evaluating A at periapsis, where r˙ is purely tangential (perpendicular to r^) with speed vp=h/rp, gives r˙×h pointing along −r^ with magnitude h2/rp, so A=(rph2−GM)r^; using rp=p/(1+e)=GM(1+e)h2 from the orbit equation gives A=GMer^p, confirming ∣A∣=GMe pointing exactly toward periapsis. ■