The title of this paper could have been ‘Geometry in five dimensions over
GF
(3)’ (cf. Edge 1954), or ‘The geometry of the second Mathieu group’, or ‘Duads and synthemes’, or ‘Hexastigms’, or simply ‘Some thoughts on the number 6’. The words actually chosen acknowledge the inspiration of the late H. F. Baker, whose last book (Baker 1946) develops the idea of duads and synthemes in a different direction. The special property of the number 6 that makes the present development possible is the existence of an outer automorphism for the symmetric group of this degree. The consequent group of order 1440 is described abstractly in §1, topologically in §2, and geometrically in §§3 to 7. The kernel of the geometrical discussion is in §5, where the chords of a non-ruled quadric in the finite projective space
PG
(3, 3) are identified with the edges of a graph having an unusually high degree of regularity (Tutte 1958). It is seen in §4 that the ten points which constitute this quadric can be derived very simply from a ‘hexastigm ’ consisting of six points in
PG
(4, 3) (cf. Coxeter 1958). The connexion with Edge’s work is described in §6. Then §7 shows that the derivation of the quadric from a hexastigm can be carried out in two distinct ways, suggesting the use of a second hexastigm in a different 4-space. It is found in §8 that the consequent configuration of twelve points in
PG
(5, 3) can be divided into two hexastigms in 66 ways. The whole set of 132 hexastigms forms a geometrical realization of the Steiner system
s
(5, 6, 12), whose group is known to be the quintuply transitive Mathieu group
M
12
, of order 95040. Finally, §9 shows how the same 5-dimensional configuration can be regarded (in 396 ways) as a pair of mutually inscribed simplexes, like Möbius’s mutually inscribed tetrahedra in ordinary space of 3 dimensions.
The Smith group and the critical group of the Grassmann graph of lines in finite projective space and of its complement
