This paper was converted on www.awesomepapers.org from LaTeX by an anonymous user.
Want to know more? Visit the Converter page.

Counterexamples to the interpolating conjecture on partial-dual genus polynomials of ribbon graphs

Qi Yan
School of Mathematics
China University of Mining and Technology
P. R. China
Xian’an Jin111Corresponding author.
School of Mathematical Sciences
Xiamen University
P. R. China
Email:[email protected]; [email protected]
Abstract

Gross, Mansour and Tucker introduced the partial-dual orientable genus polynomial and the partial-dual Euler genus polynomial. They showed that the partial-dual genus polynomial for an orientable ribbon graph is interpolating and gave an analogous conjecture: The partial-dual Euler-genus polynomial for any non-orientable ribbon graph is interpolating. In this paper, we first give some counterexamples to the conjecture. Then motivated by these counterexamples, we further find two infinite classes of counterexamples.

keywords:
Ribbon graph, partial-dual genus polynomial, conjecture, counterexample.
MSC:
05C10, 05C30, 05C31, 57M15
journal: European J. Combin.

1 Introduction

We assume that the readers are familiar with the basic knowledge of topological graph theory and in particular the ribbon graphs and partial duals, and we refer the readers to [1, 2, 3, 5]. Let GG be a ribbon graph and AE(G)A\subseteq E(G). We denote by GAG^{A} the partial dual of GG with respect to AA.

Gross, Mansour and Tucker [4] introduced the partial-dual orientable genus polynomials for orientable ribbon graphs and the partial-dual Euler genus polynomials for arbitrary ribbon graphs.

Definition 1.

[4] The partial-dual Euler genus polynomial of any ribbon graph GG is the generating function

εG(z)=AE(G)zε(GA){}^{\partial}\varepsilon_{G}(z)=\sum_{A\subseteq E(G)}z^{\varepsilon(G^{A})}

that enumerates partial duals by Euler genus. The partial-dual orientable genus polynomial of an orientable ribbon graph GG is the generating function

ΓG(z)=AE(G)zγ(GA){}^{\partial}\Gamma_{G}(z)=\sum_{A\subseteq E(G)}z^{\gamma(G^{A})}

that enumerates partial duals by orientable genus.

They posed some research problems and made some conjectures. Conjecture 5.3 in their paper states that

Conjecture 2.

[4] (Interpolating). The partial-dual Euler-genus polynomial εG(z){}^{\partial}\varepsilon_{G}(z) for any non-orientable ribbon graph GG is interpolating.

In this paper, we first give some counterexamples to Conjecture 2. Motivated by these counterexamples, we then find two infinite classes of counterexamples to the conjecture.

2 Some counterexamples

A bouquet is a ribbon graph having only one vertex. A signed rotation of a bouquet is a cyclic ordering of the half-edges at the vertex and if the edge is an untwisted loop, then we give the same sign ++ to the corresponding two half-edges, and give the different signs (one ++, the other -) otherwise. The sign ++ is always omitted. See Figure 1 for an example. Sometimes we will use the signed rotation to represent the bouquet itself. A signed rotation is called prime if it can not be cut into two parts such that the two half-edges of each edge belong to a single part. We shall only consider prime counterexamples.

Example 3.

Let BB be the bouquet with the signed rotation

(1,2,3,4,2,1,3,4)(-1,-2,3,4,2,1,3,4)

as shown in Figure 1. We have εB(z)=4z2+12z4{}^{\partial}\varepsilon_{B}(z)=4z^{2}+12z^{4} (see Table 1 for details).

Refer to caption
Figure 1: The signed rotation of the bouquet is (1,2,3,4,2,1,3,4)(-1,-2,3,4,2,1,3,4).
AA ε(A)\varepsilon(A) ε(Ac)\varepsilon(A^{c}) ε(BA)\varepsilon(B^{A})
\emptyset 0 4 4
{1}\{1\} 1 3 4
{2}\{2\} 1 3 4
{3}\{3\} 0 2 2
{4}\{4\} 0 2 2
{1,2}\{1,2\} 2 2 4
{1,3}\{1,3\} 2 2 4
{1,4}\{1,4\} 2 2 4
{2,3}\{2,3\} 2 2 4
{2,4}\{2,4\} 2 2 4
{3,4}\{3,4\} 2 2 4
{1,2,3}\{1,2,3\} 2 0 2
{1,2,4}\{1,2,4\} 2 0 2
{1,3,4}\{1,3,4\} 3 1 4
{2,3,4}\{2,3,4\} 3 1 4
{1,2,3,4}\{1,2,3,4\} 4 0 4
Table 1: Euler genera of all partial duals of (1,2,3,4,2,1,3,4)(-1,-2,3,4,2,1,3,4).

Example 3 is the first counterexample we found, and there are no counterexamples having fewer edges. We then found more counterexamples to Conjecture 2 as listed in Table 2 with the help of computer.

The signed rotation of BB εB(z){}^{\partial}\varepsilon_{B}(z)
(1,2,3,4,5,1,4,5,2,3)(-1,2,3,4,5,1,4,5,2,3) 8z2+16z4+8z58z^{2}+16z^{4}+8z^{5}
(1,2,3,1,4,2,5,4,3,5)(-1,-2,3,1,4,2,5,4,3,5) 2z+10z3+8z4+12z52z+10z^{3}+8z^{4}+12z^{5}
(1,2,3,2,4,5,6,1,5,6,3,4)(-1,2,3,2,4,5,6,1,5,6,3,4) 8z2+32z4+16z5+8z68z^{2}+32z^{4}+16z^{5}+8z^{6}
(1,2,1,3,4,5,6,2,5,6,3,4)(-1,2,1,3,4,5,6,2,5,6,3,4) 16z3+40z5+8z616z^{3}+40z^{5}+8z^{6}
(1,2,3,1,4,2,5,4,3,6,5,6)(-1,-2,3,1,4,2,5,4,3,6,5,6) 2z+14z3+12z4+28z5+8z62z+14z^{3}+12z^{4}+28z^{5}+8z^{6}
(1,2,3,4,5,6,2,1,5,6,3,4)(-1,-2,3,4,5,6,2,1,5,6,3,4) 8z2+24z4+32z68z^{2}+24z^{4}+32z^{6}
(1,2,3,2,4,3,5,6,7,1,6,7,4,5)(-1,2,3,2,4,3,5,6,7,1,6,7,4,5) 8z2+48z4+16z5+40z6+16z78z^{2}+48z^{4}+16z^{5}+40z^{6}+16z^{7}
(1,2,3,4,5,6,7,1,6,7,4,5,2,3)(-1,2,3,4,5,6,7,1,6,7,4,5,2,3) 16z2+48z4+48z6+16z716z^{2}+48z^{4}+48z^{6}+16z^{7}
(1,2,1,3,4,3,5,6,7,2,6,7,4,5)(-1,2,1,3,4,3,5,6,7,2,6,7,4,5) 16z3+80z5+16z6+16z716z^{3}+80z^{5}+16z^{6}+16z^{7}
(1,2,3,4,5,6,7,1,4,5,6,7,2,3)(-1,2,3,4,5,6,7,1,4,5,6,7,2,3) 32z4+64z6+32z732z^{4}+64z^{6}+32z^{7}
\cdots\cdots \cdots\cdots
Table 2: Other counterexamples to Conjecture 2.

3 Two infinite classes of counterexamples

In this section, motivated by examples in Table 2, we further give two infinite classes of counterexamples to Conjecture 2. The following lemma will be used.

Lemma 4.

[4] Let BB be a bouquet, and let AE(B)A\subseteq E(B). Then

ε(BA)=ε(A)+ε(Ac).\varepsilon(B^{A})=\varepsilon(A)+\varepsilon(A^{c}).

In addition, a technique called band move in knot theory [6] will be used, that is, a deformation of the bouquet by sliding one of the two ends of a ribbon along the boundary of the bouquet over other ribbons. This move does not change the number of boundary components. See Figures 3 and 4.

3.1 Infinite class 1

For each n1n\geq 1, let B2n+1B_{2n+1} be the bouquet with the signed rotation

(1,2,3,,2n,2n+1,1,2n,2n+1,,2,3),(1,2,3,\cdots,2n,2n+1,-1,2n,2n+1,\cdots,2,3),

as shown in Figure 2.

Refer to caption
Figure 2: The bouquet B2n+1B_{2n+1}.

Note that the edge 11 is interlaced with all other edges and 2i,2i+12i,2i+1 are interlaced with each other for 1in1\leq i\leq n. Let AE(B2n+1)A\subseteq E(B_{2n+1}) and 1in1\leq i\leq n. If {2i,2i+1}A\{2i,2i+1\}\subseteq A, we call {2i,2i+1}\{2i,2i+1\} double ribbons of AA. If 2iA2i\in A but 2i+1A2i+1\notin A (or 2i+1A2i+1\in A but 2iA2i\notin A), we call 2i2i (or 2i+12i+1) a single ribbon of AA.

Lemma 5.

Let AE(B2n+1)A\subseteq E(B_{2n+1}) and let s(A)s(A) be the number of single ribbons of AA. Then

ε(B2n+1A)={2n+1,whens(A)=0;2n2s(A)+2,whens(A)>0.\displaystyle\varepsilon({B_{2n+1}}^{A})=\left\{\begin{array}[]{ll}2n+1,&\mbox{when}~{}s(A)=0;\\ 2n-2s(A)+2,&\mbox{when}~{}s(A)>0.\end{array}\right.
Proof.

First observe that s(Ac)=s(A)s(A^{c})=s(A). We can assume that 1A1\in A and let f(B)f(B) denote the number of boundary components of a bouquet BB.

If the maximum labelled edge of the signed rotation of AcA^{c} appears as double ribbons, that is, Ac=(P,2i,2i+1,2i,2i+1,Q)A^{c}=(P,2i,2i+1,2i,2i+1,Q), where PP and QQ are strings, then f(Ac)=f(P,Q)f(A^{c})=f(P,Q). If the maximum labelled edge of the signed rotation of AcA^{c} appears as a single ribbon, that is, Ac=(P,2i,2i,Q)A^{c}=(P,2i,2i,Q) or Ac=(P,2i+1,2i+1,Q)A^{c}=(P,2i+1,2i+1,Q), then f(Ac)=f(P,Q)+1f(A^{c})=f(P,Q)+1. Repeating the previous argument leads to f(Ac)=s(Ac)+1=s(A)+1f(A^{c})=s(A^{c})+1=s(A)+1.

If the minimum labelled edge except 1 of the signed rotation of AA appears as a single ribbon, that is, A=(1,2j,P,1,Q,2j)A=(1,2j,P,-1,Q,2j) (or A=(1,2j+1,P,1,Q,2j+1)A=(1,2j+1,P,-1,Q,2j+1)), then f(A)=f(P,2j,1,2j,1,Q)f(A)=f(P,2j,1,2j,-1,Q) as shown in Figure 3.

Refer to caption
Figure 3: By sliding PP along the outer edge of ribbon 2j2j we change (1,2j,P,1,Q,2j)(1,2j,P,-1,Q,2j) to (P,2j,1,2j,1,Q).(P,2j,1,2j,-1,Q).

Since both PP and QQ do not contain edge 11 and s(P,Q)=s(A)1s(P,Q)=s(A)-1, it follows that

f(A)=f(P,Q)=s(P,Q)+1=s(A).f(A)=f(P,Q)=s(P,Q)+1=s(A).

If the minimum labelled edge except 1 of the signed rotation of AA appears as double ribbons, that is, A=(1,2j,2j+1,P,1,Q,2j,2j+1)A=(1,2j,2j+1,P,-1,Q,2j,2j+1), then

f(A)=f(2j,2j+1,2j,2j+1,1,P,1,Q)f(A)=f(2j,2j+1,2j,2j+1,1,P,-1,Q)

as shown in Figure 4. If s(A)=0s(A)=0, then

f(A)\displaystyle f(A) =\displaystyle= f(2j,2j+1,2j,2j+1,1,P,1,Q)\displaystyle f(2j,2j+1,2j,2j+1,1,P,-1,Q)
=\displaystyle= f(1,P,1,Q)==f(1,1)=1.\displaystyle f(1,P,-1,Q)=\cdots=f(1,-1)=1.

Otherwise, s(A)>0s(A)>0, then repeat the above process, we have f(A)=s(A).f(A)=s(A). Therefore,

f(A)={1,whens(A)=0;s(A),whens(A)>0.\displaystyle f(A)=\left\{\begin{array}[]{ll}1,&\mbox{when}~{}s(A)=0;\\ s(A),&\mbox{when}~{}s(A)>0.\end{array}\right.
Refer to caption
Figure 4: By sliding the ribbon 1 along the boundary of the handle formed by ribbons 2j2j and 2j+12j+1 we change (1,2j,2j+1,P,1,Q,2j,2j+1)(1,2j,2j+1,P,-1,Q,2j,2j+1) to (2j,2j+1,2j,2j+1,1,P,1,Q).(2j,2j+1,2j,2j+1,1,P,-1,Q).

Since ε(B2n+1A)=ε(A)+ε(Ac)\varepsilon({B_{2n+1}}^{A})=\varepsilon(A)+\varepsilon(A^{c}) by Lemma 4 and

1|A|+f(A)=2ε(A),1|Ac|+f(Ac)=2ε(Ac)1-|A|+f(A)=2-\varepsilon(A),1-|A^{c}|+f(A^{c})=2-\varepsilon(A^{c})

by the Euler formulas, it follows that

ε(B2n+1A)\displaystyle\varepsilon({B_{2n+1}}^{A}) =\displaystyle= |A|+|Ac|f(A)f(Ac)+2\displaystyle|A|+|A^{c}|-f(A)-f(A^{c})+2
=\displaystyle= {2n+1,whens(A)=0;2n2s(A)+2,whens(A)>0.\displaystyle\left\{\begin{array}[]{ll}2n+1,&\mbox{when}~{}s(A)=0;\\ 2n-2s(A)+2,&\mbox{when}~{}s(A)>0.\end{array}\right.

Theorem 6.

The partial-dual Euler genus polynomial for B2n+1B_{2n+1} is given by the formula

εB2n+1(z)=2n+1z2n+1+s=1n2n+1(ns)z2n2s+2.{}^{\partial}\varepsilon_{B_{2n+1}}(z)=2^{n+1}z^{2n+1}+\sum_{s=1}^{n}2^{n+1}\dbinom{n}{s}z^{2n-2s+2}.
Proof.

By Lemma 5, we have the following two cases:

  1. 1.

    ε(B2n+1A)=2n+1\varepsilon({B_{2n+1}}^{A})=2n+1 if and only if s(A)=0s(A)=0. Then we can choose AA in 2n+12^{n+1} ways.

  2. 2.

    ε(B2n+1A)=2n2s+2\varepsilon({B_{2n+1}}^{A})=2n-2s+2 where 1sn1\leq s\leq n if and only if s(A)=ss(A)=s. Then we can choose ss single edges for the ribbon subset AA in (ns)2s\dbinom{n}{s}2^{s} ways and then select the remaining double edges in 2ns2^{n-s} ways, and AA may or may not contain the edge 11. Hence, we can choose AA in 2n+1(ns)2^{n+1}\dbinom{n}{s} ways.

We obtain the formula. ∎

Remark 7.

By Theorem 6, B2n+1B_{2n+1} is a counterexample for each n2n\geq 2.

3.2 Infinite class 2

For each n1n\geq 1, let C2n+2C_{2n+2} be the bouquet with the signed rotation

(1,2,3,4,,2n+1,2n+2,2,1,2n+1,2n+2,,3,4)(1,2,3,4,\cdots,2n+1,2n+2,-2,-1,2n+1,2n+2,\cdots,3,4)

as shown in Figure 5.

Refer to caption
Figure 5: The bouquet C2n+2C_{2n+2}.

Note that the edges 1,21,2 are interlaced with all other edges but 1,21,2 are not interlaced with each other and 2i+1,2i+22i+1,2i+2 are interlaced with each other for 1in1\leq i\leq n. Let AE(C2n+2)A\subseteq E(C_{2n+2}) and 1in1\leq i\leq n. If {2i+1,2i+2}A\{2i+1,2i+2\}\subseteq A, we call {2i+1,2i+2}\{2i+1,2i+2\} double ribbons of AA. If 2i+1A2i+1\in A but 2i+2A2i+2\notin A (or 2i+2A2i+2\in A but 2i+1A2i+1\notin A), we call 2i+12i+1 (or 2i+22i+2) a single ribbon of AA.

Lemma 8.

Let AE(C2n+2)A\subseteq E(C_{2n+2}) and let s(A)s(A) be the number of single ribbons of AA. Then ε(C2n+2A)=\varepsilon({C_{2n+2}}^{A})=

{2n+2,if one of 1, 2 is inAand the other is inAcands(A)=0;2n2s(A)+4,if one of 1, 2 is inAand the other is inAcands(A)>0;2n2s(A)+2,if 1, 2 are both inAor both inAc.\displaystyle\left\{\begin{array}[]{ll}2n+2,&\mbox{if one of 1, 2 is in}~{}A~{}\mbox{and the other is in}~{}A^{c}~{}\mbox{and}~{}s(A)=0;\\ 2n-2s(A)+4,&\mbox{if one of 1, 2 is in}~{}A~{}\mbox{and the other is in}~{}A^{c}~{}\mbox{and}~{}s(A)>0;\\ 2n-2s(A)+2,&\mbox{if 1, 2 are both in}~{}A~{}\mbox{or both in}~{}A^{c}.\end{array}\right.
Proof.

We know that s(A)=s(Ac)s(A)=s(A^{c}). If one of 1, 2 is in AA and the other is in AcA^{c}, then, as in the proof of Lemma 5, we have

f(A)=f(Ac)={1,whens(A)=0;s(A),whens(A)>0.\displaystyle f(A)=f(A^{c})=\left\{\begin{array}[]{ll}1,&\mbox{when}~{}s(A)=0;\\ s(A),&\mbox{when}~{}s(A)>0.\end{array}\right.

If 1,2A1,2\in A, then f(Ac)=s(A)+1f(A^{c})=s(A)+1.

  1. 1.

    If the minimum labelled edge except 1, 2 of the signed rotation of AA appears as a single ribbon, that is, A=(1,2,2j+1,P,2,1,Q,2j+1)A=(1,2,2j+1,P,-2,-1,Q,2j+1) (or A=(1,2,2j+2,P,2,1,Q,2j+2)A=(1,2,2j+2,P,-2,-1,Q,2j+2)), then

    f(A)=f(P,2j+1,1,2,2j+1,2,1,Q).f(A)=f(P,2j+1,1,2,2j+1,-2,-1,Q).

    Since both PP and QQ do not contain edges 1,21,2 and s(P,Q)=s(A)1s(P,Q)=s(A)-1, it follows that

    f(A)\displaystyle f(A) =\displaystyle= f(P,Q)+f(2j+1,1,2,2j+1,2,1)1\displaystyle f(P,Q)+f(2j+1,1,2,2j+1,-2,-1)-1
    =\displaystyle= f(P,Q)+1=(s(P,Q)+1)+1=s(A)+1.\displaystyle f(P,Q)+1=(s(P,Q)+1)+1=s(A)+1.
  2. 2.

    If the minimum labelled edge except 1, 2 of the signed rotation of AA appears as double ribbons, that is,

    A=(1,2,2j+1,2j+2,P,2,1,Q,2j+1,2j+2),A=(1,2,2j+1,2j+2,P,-2,-1,Q,2j+1,2j+2),

    then

    f(A)=f(2j+1,2j+2,2j+1,2j+2,1,2,P,2,1,Q).f(A)=f(2j+1,2j+2,2j+1,2j+2,1,2,P,-2,-1,Q).

    If s(A)=0s(A)=0, then

    f(A)\displaystyle f(A) =\displaystyle= f(2j+1,2j+2,2j+1,2j+2,1,2,P,2,1,Q)\displaystyle f(2j+1,2j+2,2j+1,2j+2,1,2,P,-2,-1,Q)
    =\displaystyle= f(1,2,P,2,1,Q)==f(1,2,2,1)=1.\displaystyle f(1,2,P,-2,-1,Q)=\cdots=f(1,2,-2,-1)=1.

    Otherwise, s(A)>0s(A)>0, then repeat the above process, we have f(A)=s(A)+1.f(A)=s(A)+1. Therefore, we can see that f(A)=s(A)+1f(A)=s(A)+1 for 0s(A)n.0\leq s(A)\leq n.

Since ε(C2n+2A)=ε(A)+ε(Ac)\varepsilon({C_{2n+2}}^{A})=\varepsilon(A)+\varepsilon(A^{c}) by Lemma 4 and

1|A|+f(A)=2ε(A),1|Ac|+f(Ac)=2ε(Ac),1-|A|+f(A)=2-\varepsilon(A),1-|A^{c}|+f(A^{c})=2-\varepsilon(A^{c}),

it follows that ε(C2n+2A)=|A|+|Ac|f(A)f(Ac)+2=\varepsilon({C_{2n+2}}^{A})=|A|+|A^{c}|-f(A)-f(A^{c})+2=

{2n+2,if one of 1, 2 is inAand the other is inAcands(A)=0;2n2s(A)+4,if one of 1, 2 is inAand the other is inAcands(A)>0;2n2s(A)+2,if 1, 2 are both inAor both inAc.\displaystyle\left\{\begin{array}[]{ll}2n+2,&\mbox{if one of 1, 2 is in}~{}A~{}\mbox{and the other is in}~{}A^{c}~{}\mbox{and}~{}s(A)=0;\\ 2n-2s(A)+4,&\mbox{if one of 1, 2 is in}~{}A~{}\mbox{and the other is in}~{}A^{c}~{}\mbox{and}~{}s(A)>0;\\ 2n-2s(A)+2,&\mbox{if 1, 2 are both in}~{}A~{}\mbox{or both in}~{}A^{c}.\end{array}\right.

Theorem 9.

The partial-dual Euler genus polynomial for C2n+2C_{2n+2} is given by the formula

εC2n+2(z)=2n+1(n+2)z2n+2+s=2n2n+1((ns)+(ns1))z2n2s+4+2n+1z2.{}^{\partial}\varepsilon_{C_{2n+2}}(z)=2^{n+1}(n+2)z^{2n+2}+\sum_{s=2}^{n}2^{n+1}\left(\dbinom{n}{s}+\dbinom{n}{s-1}\right)z^{2n-2s+4}+2^{n+1}z^{2}.
Proof.

By Lemma 8, we have the following three cases:

  1. 1.

    ε(C2n+2A)=2n+2\varepsilon({C_{2n+2}}^{A})=2n+2 if and only if one of 1,21,2 is in AA and the other is in AcA^{c} and AA has no single ribbons (or only one single ribbon) or 1,21,2 are both in AA or both in AcA^{c} and AA has no single ribbons. Then we can choose AA in 2n+1+2n+1(n1)+2n+1=2n+1(n+2)2^{n+1}+2^{n+1}\dbinom{n}{1}+2^{n+1}=2^{n+1}(n+2) ways.

  2. 2.

    ε(C2n+2A)=2n2s+4\varepsilon({C_{2n+2}}^{A})=2n-2s+4 where 2sn2\leq s\leq n if and only if one of 1,21,2 is in AA and the other is in AcA^{c} and AA has ss single ribbons or 1,21,2 are both in AA or both in AcA^{c} and AA has s1s-1 single ribbons. Then we can choose AA in 2n+1((ns)+(ns1))2^{n+1}\left(\dbinom{n}{s}+\dbinom{n}{s-1}\right) ways.

  3. 3.

    ε(C2n+2A)=2\varepsilon({C_{2n+2}}^{A})=2 if and only if 1,21,2 are both in AA or both in AcA^{c} and AA has nn single ribbons. Then we can choose AA in 2n+12^{n+1} ways.

Remark 10.

By Theorem 9, C2n+2C_{2n+2} is a counterexample for each n1n\geq 1. In particular, when n=1n=1, C4C_{4} is exactly the bouquet in Example 3.

4 Acknowledgements

This work is supported by NSFC (Nos. 12171402, 12101600) and the Fundamental Research Funds for the Central Universities (Nos. 20720190062, 2021QN1037). We thank the referees sincerely for their valuable comments.

References

  • [1] B. Bollobás and O. Riordan, A polynomial of graphs on surfaces, Math. Ann. 323(1) (2002) 81–96.
  • [2] S. Chmutov, Generalized duality for graphs on surfaces and the signed Bollobás-Riordan polynomial, J. Combin. Theory Ser. B 99 (2009) 617–638.
  • [3] J. A. Ellis-Monaghan and I. Moffatt, Graphs on Surfaces, Springer New York, 2013.
  • [4] J. L. Gross, T. Mansour and T. W. Tucker, Partial duality for ribbon graphs, I: Distributions, European J. Combin. 86 (2020) 103084.
  • [5] J. L. Gross and T. W. Tucker, Topological Graph Theory, John Wiley & Sons, Inc. New York, 1987.
  • [6] C. Livingston, Knot Theory, The Mathematical Association of America, 1993.