Polyhedra without diagonals. II (Q1046884)

The author gives a new proof of a statement contained in his earlier article [Period. Math. Hungar. 15, 41--49 (1984; Zbl 0545.52004)], because the old proof has been criticized by \textit{B. Grünbaum} and \textit{G. C. Shepard} [Amer. Math. Monthly 101, No.~2, 109--128 (1994; Zbl 0807.52010)] as being incomplete. The statement to be proved is that every polyhedron in three-space with five vertices has a diagonal. The truth of the statement depends very much on the exact definition of `polyhedron' and `diagonal', which are somewhat subtle. A polyhedron by definition is a union of a finite set \(\mathcal S\) of tetrahedra forming a simplicial complex. Its edge skeleton consists of those edges of tetrahedra that belong to at least two triangles in the topological boundary of \(\bigcup \mathcal S\) which are not contained in the same plane, and its proper vertices are those vertices of tetrahedra that belong to at least two noncollinear edges in the edge skeleton. Finally, a diagonal is any segment joining two proper vertices and not contained in the union of the edge skeleton. The proof is given by a systematic scrutiny of all possible configurations.