The Markoff equation is a classical equation in number theory involving three variables. When you look at its solutions over finite fields (essentially doing arithmetic with a clock-like number system based on a prime number p), the solutions form a network called a Markoff graph. In this graph, each solution is a node, and two nodes are connected by an edge if one can be obtained from the other by a specific algebraic operation called a Vieta involution. Researchers care about these graphs because they connect number theory, geometry, and the study of how information spreads through networks.
A key question about these graphs is how well-connected they are. Prior work by Bourgain, Gamburd, and Sarnak showed that for large enough primes, the entire Markoff graph forms a single connected component, meaning you can travel between any two nodes along some path. This paper sharpens that picture by studying "vertex connectivity," which asks how many nodes you would need to remove before the graph falls apart into disconnected pieces. The authors prove that whenever the Markoff graph is connected at all, it is actually 2-connected, meaning you need to remove at least two carefully chosen nodes to disconnect it. Since the graph is known to be connected for all sufficiently large primes, it is also 2-connected for all sufficiently large primes.
The authors also show that this result is the best possible in a precise sense: for any prime of 7 or larger, the graph is never 3-connected, meaning there always exist two nodes whose removal breaks the graph apart. Together, these results give a complete and tight answer to how robust the connectivity of Markoff graphs is, and they contribute to the broader program of understanding whether these graphs form an "expander family," a property related to how efficiently information or random walks spread through the network.