On a self dual 3-sphere of Peter McMullen (Q5932714)

In his doctoral dissertation [University of Birmingham, 1968] \textit{P. McMullen} described a 3-sphere by topologically gluing sets of 5 facets of the 600-cell with a common edge to double pyramids over a 5-gon, also described geometrically as a spherical 3-honeycomb. It is self dual. But McMullen asked whether it is polytopal. The authors here look at its automorphism group and its relation to the Coxeter group \(H_4\) of the 600-cell. Another equifacetted polyhedral 3-sphere (240-cell) with 240 facets and 120 vertices has the same automorphism group. Both of these 3-spheres and the polar dual of the second cannot occur as a boundary complex of a convex 4-polytope with the automorphism group as their full Euclidean symmetry. Whether any of these three 4-polytopes exists is not known. The methods of computational synthetic geometry fail. The authors show that even a certain subgroup of order 5 of the full combinatorial symmetry group of order 1200 cannot occur as a Euclidean symmetry for McMullen's original ``polytope''.