London’s spanning-tree extremal conjecture for C4-free graphs

Determine whether, for every prime power q and n=q^2+q+1, the maximum number of spanning trees among n-vertex C4-free graphs equals n^{(n-3)/2}, with equality attained precisely by the orthogonal polarity graphs.

Background

For every prime power q, the Erdős–Rényi orthogonal polarity graph ER_q is C4-free and has n{(n-3)/2} spanning trees, providing a general lower bound for the extremal quantity st(n,C4). London conjectured that this lower bound is always sharp and that the only extremal graphs are orthogonal polarity graphs.

The paper proves the conjecture for every prime power q outside the exceptional set {7,9,11,13}. For the four exceptional values, the uniform theorem establishes the desired bound for graphs with at most m_q=\frac12q(q+1)2 edges, but does not settle graphs with more than m_q edges; consequently, the conjecture remains unresolved in those cases.

References

London conjectured that this lower bound is best possible and that the extremal graphs are precisely the orthogonal polarity graphs.

— A Turán-type extremal problem for the number of spanning trees in $C_4$-free graphs  (2609.34616 - Xu et al., 28 Sep 2026) in Section 1, immediately before Conjecture 1 (Conjecture 1); Section 6

It would therefore be interesting to determine whether the spanning tree bound can be proved for these four cases without first resolving the corresponding extremal edge problem for $C_4$.

— A Turán-type extremal problem for the number of spanning trees in $C_4$-free graphs  (2609.34616 - Xu et al., 28 Sep 2026) in Section 6, Concluding remarks

More precisely, given a fixed graph $F$, one may ask which $n$-vertex $F$-free graphs maximize the number of spanning trees, and whether the extremal graphs for the classical Turán problem for $F$ also maximize the number of spanning trees.

— A Turán-type extremal problem for the number of spanning trees in $C_4$-free graphs  (2609.34616 - Xu et al., 28 Sep 2026) in Section 6, Concluding remarks