| Configuration | Vertices | Edges | Faces | Cells | Result | |||
| {5, 3, 3} | 600 | 7200 | {}@18 | 7200 | {3} | 3600 | {3, 3} | 720 * {3, 3, 3} |
| 1200 | 1200 | {3} | 600 | {3, 3} | 120 * {3, 3, 3} | |||
Because the edges have different graph distances in this section, the symmetry group splits them in two different orbits, and that splitting continues until the end. At the end we find in one orbit a compound of 720 {3, 3, 3}'s while in the other orbit we find a compound of 120 {3, 3, 3}'s. Interestingly, Coxeter notes only the second compound and appears not to have seen the first compound. I do not know (yet) whether the first compound can be split in subcompounds where all vertices are covered, at least the rotation group also does not give information about that. But it is extremely likely that there are quite a few variants. And possibly there are also variants with mixed kinds of edges.