S. Stahl conjectured that the zeros of genus polynomials are real. In this note, we disprove this conjecture.

For every polytope $\mathcal{P}$ there is the universal regular polytope of the same rank as $\mathcal{P}$ corresponding to the Coxeter group $\mathcal{C} =[\infty, \dots, \infty]$. For a given automorphism $d$ of $\mathcal{C}$, using monodromy groups, we construct a combinatorial structure $\mathcal{P}^d$. When $\mathcal{P}^d$ is a polytope isomorphic to $\mathcal{P}$ we say that $\mathcal{P}$ is self-invariant with respect to $d$, or $d$-invariant. We develop algebraic tools for investigating these operations on polytopes, and in particular give a criterion on the existence of a $d$\nobreakdash-auto\-morphism of a given order. As an application, we analyze properties of self-dual edge-transitive polyhedra and polyhedra with two flag-orbits. We investigate properties of medials of such polyhedra. Furthermore, we give an example of a self-dual equivelar polyhedron which contains no polarity (duality of order 2). We also extend the concept of Petrie dual to higher dimensions, and we show how it can be dealt with using self-invariance.

S. Stahl (Canad. J. Math. \textbf{49}(1997), no. 3, 617--640) conjectured that the zeros of genus polynomial are real. L. Liu and Y. Wang disproved this conjecture on the basis of Example 6.7. In this note, it is pointed out that there is an error in this example and a new generating matrix and initial vector are provided.

No abstract.

In the genus polynomial of the graph $G$, the coefficient of $x^k$ is the number of distinct embeddings of the graph $G$ on the oriented surface of genus $k$. It is shown that for several infinite families of graphs all the zeros of the genus polynomial are real and negative. This implies that their coefficients, which constitute the genus distribution of the graph, are log concave and therefore also unimodal. The geometric distribution of the zeros of some of these polynomials is also investigated and some new genus polynomials are presented.

In the present paper, we face the problem of classifying classes of orientable PL $5$-manifolds $M^5$ with $h \geq 1$ boundary components, by making use of a combinatorial invariant called {\it regular genus} ${\cal G}(M^5)$. In particular, a complete classification up to regular genus five is obtained: $${\cal G}(M^5) = \gG \leq 5 \Longrightarrow M^5 \cong \#_{\varrho - \gbG}(\bdo) \# \smo_{\gbG},$$ where $\gbG = {\cal G}(\partial M^5)$ denotes the regular genus of the boundary $\partial M^5$ and $\smo_{\gbG}$ denotes the connected sum of $h\geq 1$ orientable $5$-dimensional handlebodies $\cmo_{\alpha_i}$ of genus $\alpha_i\geq 0$ ($i=1,\ldots, h$), so that $\sum_{i=1}^h \alpha_i = \gbG.$ \par Moreover, we give the following characterizations of orientable PL $5$-manifolds $M^5$ with boundary satisfying particular conditions related to the ``gap'' between ${\cal G}(M^5)$ and either ${\cal G}(\partial M^5)$ or the rank of their fundamental group $\rk\bigl(\pi_1(M^5)\bigr)$: $$\displaylines{{\cal G}(\partial M^5)= {\cal G}(M^5) = \varrho \Longleftrightarrow M^5 \cong \smo_{\gG}\cr {\cal G}(\partial M^5)= \gbG = {\cal G}(M^5)-1 \Longleftrightarrow M^5 \cong (\bdo) \# \smo_{\gbG}\cr {\cal G}(\partial M^5)= \gbG = {\cal G}(M^5)-2 \Longleftrightarrow M^5 \cong \#_2 (\bdo) \# \smo_{\gbG}\cr {\cal G}(M^5) = \rk\bigl(\pi_1(M^5)\bigr)= \varrho \Longleftrightarrow M^5 \cong \#_{\gG - \gbG}(\bdo) \# \smo_{\gbG}.\cr}$$ \par Further, the paper explains how the above results (together with other known properties of regular genus of PL manifolds) may lead to a combinatorial approach to $3$-dimensional Poincar\'e Conjecture.