A generalization of some of Folkman's constructions (see (1967) J. Comb. Theory, 3, 215-232) of the so-called semisymmetric graphs, that is regular graphs which are edge-but not vertex-transitive, was given by Marušič and Potočnik (2001, Europ. J. Combinatorics, 22, 333-349) together with a natural connection between graphs admitting 1 2 -arc-transitive group actions and certain graphs admitting semisymmetric group actions. This connection is studied in more detail in this paper. Among others, a sufficient condition for the semisymmetry of the so-called generalized Folkman graphs arising from certain graphs admitting a 1 2 -arc-transitive group action is given. Furthermore, the concepts of alter-sequence and alter-exponent is introduced and studied in great detail and then used to study the interplay of three classes of graphs: cubic graphs admitting a one-regular group action, the corresponding line graphs which admit a 1 2 -arc-transitive action of the same group and the associated generalized Folkman graphs. At the end an open problem is posed, suggesting an in-depth analysis of the structure of tetravalent 1 2 -arc-transitive graphs with alter-exponent 2.c 2002 Elsevier Science Ltd. All rights reserved.
INTRODUCTORY REMARKS A digraph D = (V, A) consists of a finite set of vertices V (D) = V and a set of arcs A(D)For an arc (u, v) of a digraph D we say that u and v are the vertices of (u, v), more pecisely, u is the tail and v is the head of (u, v). Also, we say that v is an out-neighbour of u and that u is an in-neighbour of v. , we say that D is a graph. An edge of a graph is an unordered pair {u, v} (also denoted by uv) such that (u, v) is an arc of the graph. The set of edges of the graph X is denoted by E(X ).We refer the reader to [9,25,28] for group-theoretic concepts not defined here. Let X be a graph and G a subgroup of the automorphism group AutX of X . We say that X is G-vertex-transitive, G-edge-transitive and G-arc-transitive if G acts transitively on V (X ), E(X ) and A(X ), respectively. Furthermore, X is said to be (G, Let G be a subgroup of AutX such that X is G-edge-transitive but not G-vertex-transitive. Then X is necessarily bipartite, where the two parts of the bipartition are orbits of G. Clearly † Corresponding address: IMFM,