The universal embedding dimension of the near polygon on the \(1\)-factors of a complete graph (Q1963149)

Given a point-line geometry with lines of size 3, its universal embedding over \({\mathcal F}_2\) is the \({\mathcal F}_2\)-vector space \(U = V/W\), where \(V\) is the vector space that has the point set of the geometry as basis, and \(W\) is the subspace generated by the elements \(x + y + z\) for each line \(\{x, y, z\}\). The universal embedding dimension of the geometry is \(\dim U\). A near polygon is a connected partial linear space such that for each point \(x\) and each line \(L\) there is a unique point on \(L\) closest to \(x\) in the collinearity graph. The near polygon \(H_{n - 1}\) has as its points the 1-factors of the complete graph \(K_{2n}\) and as its lines (of size 3) the sets of 1-factors on a partial 1-factor of size \(n - 2\). This near polygon has \(v = (2n - 1) (2n - 3) \ldots (3) (1)\) points and \(t + 1 = {{n}\choose{2}}\) lines on each point. In the paper under review the authors show that the universal embedding dimension of \(H_{n - 1}\) is our old friend, the Catalan number \(C_n = \frac{1}{n + 1} {{2n}\choose {n}}\). This is done by explicitly exhibiting a basis of its universal embedding, namely the 1-factors of \(K_{2n}\) without crossing edges, when the \(2n\) vertices are drawn as the vertices of a convex \(2n-gon\) in the plane. It is well known that the number of such 1-factors is \(C_{n}\). Finally, it is shown that the universal embedding dimension of \(H_{n - 1}\) equals the multiplicity of the eigenvalue \(-t - 1\) of the adjacency matrix of its collinearity graph. The authors point out that \(H_{n - 1}\) is a ``large'' subgeometry of the dual symplectic near polygon \(DSp (2n - 2, 2)\), and it is hoped that knowledge of the universal embedding dimension of \(H_{n - 1}\) will help settle the old conjecture that the universal embedding dimension of \(DSp (2d, 2)\) is \((2^d + 1)(2^{d - 1} + 1)/3\).