Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.
Updated
Statement
Printed pp. 25--26, Section 1 ("The 7-cube minus a Hamming code"), read on the page images of the version of record named on the source card. The section has no theorem label; its results are the numbered properties (1) to (6) on p. 26, and this page is named for the section.
Setting (p. 25). The vertices of the hypercube on a set are the subsets of , a vector space over under symmetric difference , with when . Take and let consist of , the seven sets (), and the complements of these eight sets. The paper states that is a subspace and a perfect -error-correcting code, hence a Hamming code: its vertices are pairwise nonadjacent and every vertex outside has exactly one neighbor in . The subgraph of the -cube induced by has vertices, is regular of valency , and is vertex-transitive.
The coloring (pp. 25--26). For an edge of with of odd weight, define by and ; the edge is red when and white otherwise. and are the red and white subgraphs, both on the vertex set .
Properties (p. 26). The paper states, with short indications for (1) to (3) and (6):
- ; for odd-weight , is an isomorphism.
- " is solvable of order 168, acts (sharply) transitively on the edges of both and , and has two orbits on their vertex set ." The group is generated by the translations by members of the even-weight subcode of , the cyclic shifts of , and the permutation of .
- " has diameter 8; for any vertex of odd weight there is a unique antipode at distance 8, where is determined by ; no two vertices of even weight have distance 8."
- Every quadrangle of has three edges of one color and one of the other. If is a white edge, then and are at distance in , joined there by a unique path .
- " has girth 10."
- " is an 8-cover of the Heawood graph, the point-line incidence graph of the Fano plane." Identifying vertices of that differ by an element of gives a graph isomorphic to the Heawood graph on , with if and only if .
So and are isomorphic cubic graphs of girth on vertices that are edge-transitive but not vertex-transitive, as the abstract (p. 25) puts it.
The three-coloring (p. 25). The paper notes that once both color classes have girth , "it follows that the edges of can be colored with three colors such that there are no monochromatic -gons for ." Section 3 (p. 28) recalls this as a three-coloring of the edges of the -cube without monochromatic quadrangle or hexagon for .
Attribution and identification (pp. 26 and 28). The paper says the graph was constructed in Dejter and Guan (its reference [7]) and may be the graph R. M. Foster constructed according to Bouwer (reference [2]). A remark added in proof (p. 28) states that it differs from the unique trivalent graph on vertices with girth in Foster's census, since it is not vertex-transitive.
Source. A. E. Brouwer, I. J. Dejter and C. Thomassen, Highly symmetric subgraphs of hypercubes, J. Algebraic Combin. 2 (1993), 25--29, doi:10.1023/A:1022472513494; Section 1 on printed pp. 25--26, the remark added in proof on p. 28.
Read depth. Claims checked: the setting, the coloring and properties (1) to (6) were read clause by clause on the page images of printed pp. 25--26. The paper's indications of proof were read but not checked; properties (4) to (6) are stated without proof apart from the identification in (6).
Proof pointer
Pp. 25--26. Property (1) is witnessed by translation by an odd-weight codeword; for (2) the paper exhibits the group of order 168 and shows it is sharply edge-transitive, preserving the parity of the weight; (3) is verified by growing the distance classes from and , which the paper says also shows that the full automorphism group is no larger than the group of (2); (6) is the quotient by . Not checked here.
Dependencies
Self-contained; the facts that is a perfect code and a subspace are stated in the paper as standard.
Bears on
No Erdős problem directly. The three-coloring it yields for is context for Section 3, whose four-coloring the paper uses against Erdős's hexagon conjecture.