Weakly Pancyclic Vertices in Graphs

• Weakly pancyclic vertices are those that lie on cycles of every length from a graph's girth to its circumference, highlighting a local cycle property.
• Research in dense nonbipartite graphs shows that beyond the BT(n) exception, at least three vertices exhibit weak pancyclicity when edge density exceeds a precise threshold.
• In connected, locally isometric graphs with maximum degree up to 6, every vertex is weakly pancyclic, underlining the importance of local structural constraints.

A vertex of a finite simple graph is termed weakly pancyclic if it is incident to cycles of every length ranging from the graph’s girth to its circumference. Weakly pancyclic vertices constitute a local strengthening of the traditional global notion of a weakly pancyclic graph. Research has focused on their appearance in dense nonbipartite graphs, locally isometric graphs, and related extremal constructions, with special attention paid to structural exceptions and optimal lower bounds on their frequency.

1. Formal Definition and Fundamental Properties

Let $G$ be a finite simple graph with vertex set $V(G)$ and edge set $E(G)$. Define the girth $g(G)$ as the length of the shortest cycle in $G$, and the circumference $c(G)$ as the length of the longest cycle.

A vertex $v \in V(G)$ is weakly pancyclic if for every integer $\ell$ satisfying $g(G)\leq \ell \leq c(G)$, there exists a cycle of length $\ell$ containing $v$: $$\forall \ell \in \{g(G), g(G)+1, \dots, c(G)\},~ \exists~ \ell\text{-cycle}~ C \subseteq G \text{ with } v \in V(C)$$ This notion is strictly local, in contrast to the property that the whole graph is weakly pancyclic (i.e., contains cycles of all lengths between girth and circumference, regardless of which vertices are contained in them) (Tang et al., 22 Jan 2026, Borchert et al., 2015).

2. Main Theoretical Results in Dense Nonbipartite Graphs

In the context of extremal graph theory, a definitive characterization for the presence of weakly pancyclic vertices in nonbipartite graphs of high edge density was established by Tang and Zhan. The principal result:

Let $G$ be a nonbipartite graph of order $n\ge5$ and size

$$|E(G)| \ge \Big\lfloor\frac{(n-1)^2}{4}\Big\rfloor + 2$$

Then one of the following holds:

• (i) $G \cong BT(n)$, where $BT(n)$ is the “Brandt–Thomason” exceptional graph (a complete bipartite graph $K_{\lfloor (n-1)/2 \rfloor,\lceil (n-1)/2 \rceil}$ with one edge identified with an edge of a triangle $C_3$), or
• (ii) $G$ contains three distinct weakly pancyclic vertices.

This result is sharp with respect to both the edge threshold and the number 3: in $BT(n)$, there are exactly two weakly pancyclic vertices, regardless of the parity of $n$, showing the necessity of the $BT(n)$ exception (Tang et al., 22 Jan 2026).

3. Vertex-Weak Pancyclicity in Locally Isometric Graphs

A graph $G$ is locally isometric if the subgraph induced by the neighborhood of every vertex is an isometric subgraph. In this setting, Borchert, Nicol, and Oellermann proved:

Let $G$ be a connected, locally isometric graph of order $n\geq \Delta+1$ and maximum degree $\Delta \leq 6$. Then every vertex $v \in V(G)$ is weakly pancyclic; that is, $v$ lies on a cycle of every length in $[g(G), c(G)]$.

For $\Delta\leq 5$, every non-fully-cycle-extendable locally isometric graph is structurally characterized (as “singly or doubly shuttered highrise” graphs), and all such cases are also vertex-weakly-pancyclic. For $\Delta=6$, two exceptions arise: either a pair of true twins of degree 6 or the graph $K_{2,4} + K_1$; in both, every vertex is weakly pancyclic (Borchert et al., 2015).

4. Structure of Extremal Examples and Sharpness of Main Theorems

Detailed extremal constructions serve both to demonstrate the sharpness of general results and to provide insight into what structural features prevent further improvement. Notably:

• The exceptional $BT(n)$ realizes the lower bound in dense nonbipartite graphs, with precisely two weakly pancyclic vertices.
• Tang and Zhan describe for each $n\ge6$ a family $G_n$ (built by augmenting $BT(n-1)$ with a new vertex and connecting it appropriately) where exactly three vertices (the triangle vertices) are weakly pancyclic—proving the “three” in the main result is optimal (Tang et al., 22 Jan 2026).
• In locally isometric graphs with $\Delta\leq6$, the only exceptions to full cycle-extendability (yet not to vertex-weak pancyclicity) are the “shuttered highrise” graphs and $K_{2,4} + K_1$ (Borchert et al., 2015).

Extremal Family	Max # w.p. vertices	Graph Invariant (order, edges)	Reference
$BT(n)$	2	$n, \left\lfloor (n-1)^2/4\right\rfloor+2$	(Tang et al., 22 Jan 2026)
$G_n$	3	$n, \left\lfloor (n-1)^2/4\right\rfloor+2$	(Tang et al., 22 Jan 2026)
$K_{2,4} + K_1$	all vertices	$7, 9$	(Borchert et al., 2015)
Singly/Doubly Shuttered Highrise	all vertices	Varies	(Borchert et al., 2015)

5. Outline of Proof Techniques and Structural Lemmas

Theorems regarding weakly pancyclic vertices employ induction on the order $n$ and distinguish between Hamiltonian and non-Hamiltonian cases:

• For Hamiltonian graphs not isomorphic to $BT(n)$, one obtains pancyclic vertices by inductively removing small-degree vertices or via neighbor-count arguments for “big” vertices. A central ingredient is Brandt's Lemma, which concerns the construction of long paths in nearly complete bipartite subgraphs, supporting the existence of required cycles of each length for candidate vertices (Tang et al., 22 Jan 2026).
• For non-Hamiltonian graphs, an extremal argument identifies a small vertex outside a longest cycle; deleting this vertex and applying the hypothesis ensures the persistence of weakly pancyclic vertices upon reintroduction (Tang et al., 22 Jan 2026).
• In locally isometric graphs, a suite of lemmas forbids sparse neighborhoods and prescribes specific local configurations, allowing for the preservation and extension of cycles through deletions, ultimately guaranteeing vertex-weak pancyclicity even in graphs with obstructions to full cycle-extendability (Borchert et al., 2015).

6. Comparison with Historical Results and Strengthenings

Brandt (1997) established that every nonbipartite graph on $n$ vertices and at least $\lfloor (n-1)^2/4 \rfloor + 2$ edges is weakly pancyclic as a whole. Tang and Zhan’s result strengthens this: not only does every such graph (excluding $BT(n)$) have all cycle lengths, but at least three vertices participate in cycles of every possible length between girth and circumference. This advances the local perspective on extremal cycle theory and delineates a clear improvement over purely global results (Tang et al., 22 Jan 2026).

7. Related Open Problems and Directions

Outstanding problems in the study of weakly pancyclic vertices include:

• Pancyclic edges in Hamiltonian graphs: Zhan (2025) conjectures that if $G$ is Hamiltonian, nonbipartite, of order $n\ge7$, size at least $\lfloor (n-1)^2/4 \rfloor + 2$, and $G\neq BT(n)$ for odd $n$, then there exists a pancyclic edge (i.e., both endpoints are pancyclic vertices) (Tang et al., 22 Jan 2026).
• Function $f(n)$ for minimum edge count: Define

$$f(n) = \min \left\{ k : \text{every nonbipartite graph of order } n,\, |E(G)|\geq k,\, \text{has a weakly pancyclic vertex} \right\}$$

It is known that $f(n)\leq \lfloor (n-1)^2/4 \rfloor + 2$, with explicit values for small $n$ provided by computation. Tight determination of $f(n)$ in general remains unresolved (Tang et al., 22 Jan 2026).

8. Broader Classes and Vertex-Weak Pancyclicity

Beyond dense nonbipartite graphs, vertex-weak pancyclicity is ensured for all vertices in connected, locally isometric graphs with $\Delta \leq 6$. Such results encompass not only fully cycle-extendable graphs but also certain structured exceptions, indicating that local geometric constraints can suffice to guarantee strong local cycle coverage even where global cycle extension properties may fail (Borchert et al., 2015).

Research in this area continues to probe structural constraints, generalizations to sparser classes, relationships with cycle extendibility, and finer-grained extremal parameters for weakly pancyclicity in both vertex and edge terms.