The study of algebraic curves over finite fields is a central topic in mathematics, with connections to coding theory and cryptography. A "maximal curve" is a special type of curve defined over a finite field that achieves the largest possible number of rational points allowed by a classical bound called the Hasse-Weil bound. The most famous example is the Hermitian curve, which is in many ways the most symmetric and well-behaved maximal curve. This paper asks a natural question about symmetry: when can the automorphism group of a maximal curve act transitively on its rational points? Transitivity here means that the symmetries of the curve can move any rational point to any other rational point, so no point is more "special" than another from the perspective of the curve's symmetries.
The authors provide a complete answer to this question for odd prime powers. The main result is that transitivity is extremely rare. If the field size involves a prime power with exponent greater than one, the only curve with this property is the Hermitian curve itself. If the exponent equals one (meaning the field size is simply a prime), there is exactly one additional example: a specific genus-three curve over the field with 25 elements, which turns out to be a particular version of the famous Klein quartic, a highly symmetric curve that has appeared throughout mathematics for over a century.
The proof strategy divides into two fundamentally different cases depending on how the automorphism group interacts with the characteristic of the field. In the "tame" case, where the prime does not divide the size of certain symmetry groups, the authors use classical tools like genus bounds and the Riemann-Hurwitz formula, eventually reducing everything to the Klein quartic. In the "wild" case, where the prime does divide these group sizes, the geometry of rational points connects directly to the internal structure of the automorphism group itself, specifically its Sylow subgroups, and the authors use a combination of finite group theory and ramification arguments to rule out all other possibilities. Together with a previously known result handling characteristic two, this gives a complete classification across all finite fields.