\(f\)-vectors of triangulated balls (Q644825)

The \textit{\(f\)-vector} of a simplicial complex \(\Delta\) records the number of \(k\)-dimensional faces of \(\Delta\), for each \(k\). The famous \textit{\(g\)-theorem}, conjectured by McMullen and proved by Billera, Lee, and Stanley, characterizes \(f\)-vectors of simplicial polytopes. \textit{L. J. Billera} and \textit{C. W. Lee}, [Eur. J. Comb. 2, 307--322 (1981; Zbl 0484.52008)] established conditions for \(f\)-vectors of triangulated balls and conjectured that these conditions are also sufficient. The paper under review presents two methods for showing that a vector is \textit{not} the \(f\)-vector of a triangulated ball. In both approaches, the author assumes that a ball with a certain \(f\)-vector exists and then shows how to split the ball along a codimension \(-1\) face into two balls, at least one of which cannot exist. As a consequence, the author proves that the Billera--Lee conditions are not sufficient in dimension \(\geq 5\). The paper also presents a construction of balls with prescribed \(f\)-vectors, which yields all possible \(f\)-vectors of balls in dimension 3 and 4. (\textit{C. W. Lee} and \textit{L. Schmidt} [Rocky Mt. J. Math. 41, No. 6, 1939--1961 (2011; Zbl 1238.52003)] have given a construction with the same result.) The author conjecturs that his construction also gives all possible \(f\)-vectors in dimension~\(5\).