c 2009 Society for Industrial and Applied Mathematics

Similar documents
MAT3707. Tutorial letter 201/1/2017 DISCRETE MATHEMATICS: COMBINATORICS. Semester 1. Department of Mathematical Sciences MAT3707/201/1/2017

(4, 2)-choosability of planar graphs with forbidden structures

Cycles and Simple Cycles. Paths and Simple Paths. Trees. Problem: There is No Completely Standard Terminology!

Complete Solutions for MATH 3012 Quiz 2, October 25, 2011, WTT

learning objectives learn what graphs are in mathematical terms learn how to represent graphs in computers learn about typical graph algorithms

Grade 7/8 Math Circles March 4/5, Graph Theory I- Solutions

16.unified Introduction to Computers and Programming. SOLUTIONS to Examination 4/30/04 9:05am - 10:00am

The University of Sydney MATH 2009

1 Introduction to Modulo 7 Arithmetic

The University of Sydney MATH2969/2069. Graph Theory Tutorial 5 (Week 12) Solutions 2008

, each of which is a tree, and whose roots r 1. , respectively, are children of r. Data Structures & File Management

d e c b a d c b a d e c b a a c a d c c e b

Paths. Connectivity. Euler and Hamilton Paths. Planar graphs.

0.1. Exercise 1: the distances between four points in a graph

Outline. 1 Introduction. 2 Min-Cost Spanning Trees. 4 Example

Present state Next state Q + M N

Lecture 20: Minimum Spanning Trees (CLRS 23)

Tangram Fractions Overview: Students will analyze standard and nonstandard

b. How many ternary words of length 23 with eight 0 s, nine 1 s and six 2 s?

QUESTIONS BEGIN HERE!

Weighted Graphs. Weighted graphs may be either directed or undirected.

An undirected graph G = (V, E) V a set of vertices E a set of unordered edges (v,w) where v, w in V

OpenMx Matrices and Operators

Algorithmic and NP-Completeness Aspects of a Total Lict Domination Number of a Graph

Math 61 : Discrete Structures Final Exam Instructor: Ciprian Manolescu. You have 180 minutes.

Graph Algorithms and Combinatorial Optimization Presenters: Benjamin Ferrell and K. Alex Mills May 7th, 2014

Planar Upward Drawings

CS 103 BFS Alorithm. Mark Redekopp

DFA Minimization. DFA minimization: the idea. Not in Sipser. Background: Questions: Assignments: Previously: Today: Then:

Outline. Computer Science 331. Computation of Min-Cost Spanning Trees. Costs of Spanning Trees in Weighted Graphs

Steinberg s Conjecture is false

QUESTIONS BEGIN HERE!

Math 166 Week in Review 2 Sections 1.1b, 1.2, 1.3, & 1.4

ECE COMBINATIONAL BUILDING BLOCKS - INVEST 13 DECODERS AND ENCODERS

5/1/2018. Huffman Coding Trees. Huffman Coding Trees. Huffman Coding Trees. Huffman Coding Trees. Huffman Coding Trees. Huffman Coding Trees

(2) If we multiplied a row of B by λ, then the value is also multiplied by λ(here lambda could be 0). namely

Constructive Geometric Constraint Solving

MCS100. One can begin to reason only when a clear picture has been formed in the imagination.

Graph Isomorphism. Graphs - II. Cayley s Formula. Planar Graphs. Outline. Is K 5 planar? The number of labeled trees on n nodes is n n-2

CSE 373: More on graphs; DFS and BFS. Michael Lee Wednesday, Feb 14, 2018

12/3/12. Outline. Part 10. Graphs. Circuits. Euler paths/circuits. Euler s bridge problem (Bridges of Konigsberg Problem)

5/9/13. Part 10. Graphs. Outline. Circuits. Introduction Terminology Implementing Graphs

Depth First Search. Yufei Tao. Department of Computer Science and Engineering Chinese University of Hong Kong

COMPLEXITY OF COUNTING PLANAR TILINGS BY TWO BARS

CS September 2018

Trees as operads. Lecture A formalism of trees

CMPS 2200 Fall Graphs. Carola Wenk. Slides courtesy of Charles Leiserson with changes and additions by Carola Wenk

VLSI Testing Assignment 2

Seven-Segment Display Driver

Exam 1 Solution. CS 542 Advanced Data Structures and Algorithms 2/14/2013

Improving Union. Implementation. Union-by-size Code. Union-by-Size Find Analysis. Path Compression! Improving Find find(e)

Solutions for HW11. Exercise 34. (a) Use the recurrence relation t(g) = t(g e) + t(g/e) to count the number of spanning trees of v 1

Graphs. Graphs. Graphs: Basic Terminology. Directed Graphs. Dr Papalaskari 1

Designing A Concrete Arch Bridge

In which direction do compass needles always align? Why?

Divided. diamonds. Mimic the look of facets in a bracelet that s deceptively deep RIGHT-ANGLE WEAVE. designed by Peggy Brinkman Matteliano

STRUCTURAL GENERAL NOTES

4.1 Interval Scheduling. Chapter 4. Greedy Algorithms. Interval Scheduling: Greedy Algorithms. Interval Scheduling. Interval scheduling.

Solutions to Homework 5

CSE 373. Graphs 1: Concepts, Depth/Breadth-First Search reading: Weiss Ch. 9. slides created by Marty Stepp

Why the Junction Tree Algorithm? The Junction Tree Algorithm. Clique Potential Representation. Overview. Chris Williams 1.

EE1000 Project 4 Digital Volt Meter

Combinatorial Optimization

CMSC 451: Lecture 4 Bridges and 2-Edge Connectivity Thursday, Sep 7, 2017

Module graph.py. 1 Introduction. 2 Graph basics. 3 Module graph.py. 3.1 Objects. CS 231 Naomi Nishimura

Slide-and-swap permutation groups. Onyebuchi Ekenta, Han Gil Jang and Jacob A. Siehler. (Communicated by Joseph A. Gallian)

CS150 Sp 98 R. Newton & K. Pister 1

BASIC CAGE DETAILS SHOWN 3D MODEL: PSM ASY INNER WALL TABS ARE COINED OVER BASE AND COVER FOR RIGIDITY SPRING FINGERS CLOSED TOP

Outline. Binary Tree

Minimum Spanning Trees

Edge-Triggered D Flip-flop. Formal Analysis. Fundamental-Mode Sequential Circuits. D latch: How do flip-flops work?

CS 241 Analysis of Algorithms

Revisiting Decomposition Analysis of Geometric Constraint Graphs

Garnir Polynomial and their Properties

Outline. Circuits. Euler paths/circuits 4/25/12. Part 10. Graphs. Euler s bridge problem (Bridges of Konigsberg Problem)

CS 461, Lecture 17. Today s Outline. Example Run

V={A,B,C,D,E} E={ (A,D),(A,E),(B,D), (B,E),(C,D),(C,E)}

Having a glimpse of some of the possibilities for solutions of linear systems, we move to methods of finding these solutions. The basic idea we shall

BASIC CAGE DETAILS D C SHOWN CLOSED TOP SPRING FINGERS INNER WALL TABS ARE COINED OVER BASE AND COVER FOR RIGIDITY

A 4-state solution to the Firing Squad Synchronization Problem based on hybrid rule 60 and 102 cellular automata

Graphs. CSC 1300 Discrete Structures Villanova University. Villanova CSC Dr Papalaskari

CSE 373: AVL trees. Warmup: Warmup. Interlude: Exploring the balance invariant. AVL Trees: Invariants. AVL tree invariants review

24.1 Sex-Linked Inheritance. Chapter 24 Chromosomal Basis of Inheritance Sex-Linked Inheritance Sex-Linked Inheritance

(Minimum) Spanning Trees

S i m p l i f y i n g A l g e b r a SIMPLIFYING ALGEBRA.

Spanning Trees. BFS, DFS spanning tree Minimum spanning tree. March 28, 2018 Cinda Heeren / Geoffrey Tien 1

(4, 2)-choosability of planar graphs with forbidden structures

V={A,B,C,D,E} E={ (A,D),(A,E),(B,D), (B,E),(C,D),(C,E)}

Physics 222 Midterm, Form: A

5/7/13. Part 10. Graphs. Theorem Theorem Graphs Describing Precedence. Outline. Theorem 10-1: The Handshaking Theorem

Walk Like a Mathematician Learning Task:

Proof of Pumping Lemma. PL Use. Example. Since there are only n different states, two of q 0, must be the same say q i

Problem solving by search

Continuous Flattening of Convex Polyhedra

Using the Printable Sticker Function. Using the Edit Screen. Computer. Tablet. ScanNCutCanvas

CSC Design and Analysis of Algorithms. Example: Change-Making Problem

# 1 ' 10 ' 100. Decimal point = 4 hundred. = 6 tens (or sixty) = 5 ones (or five) = 2 tenths. = 7 hundredths.

24CKT POLARIZATION OPTIONS SHOWN BELOW ARE REPRESENTATIVE FOR 16 AND 20CKT

Indices. Indices. Curriculum Ready ACMNA: 209, 210, 212,

Two Approaches to Analyzing the Permutations of the 15 Puzzle

Transcription:

SIAM J. DISCRETE MATH. Vol. 0, No. 0, pp. 000 000 2009 Soity or Inustril n Appli Mtmtis THE TWO-COLORING NUMBER AND DEGENERATE COLORINGS OF PLANAR GRAPHS HAL KIERSTEAD, BOJAN MOHAR, SIMON ŠPACAPAN, DAQING YANG, AND XUDING ZHU Astrt. T two-olorin numr o rps, wi ws oriinlly introu in t stuy o t m romti numr, lso ivs n uppr oun on t nrt romti numr s introu y Boroin. It is prov tt t two-olorin numr o ny plnr rp is t most nin. As onsqun, t nrt list romti numr o ny plnr rp is t most nin. It is lso sown tt t nrt ionl romti numr is t most 11 n t nrt ionl list romti numr is t most 12 or ll plnr rps. Ky wors. two-olorin numr, nrt olorin, plnr rp AMS sujt lssiition. 05C15 DOI. 10.1137/070703697 1. Introution. T two-olorin numr o rp ws introu y Cn n Slp [6] in t stuy o Rmsy proprtis o rps n us ltr in t stuy o t m romti numr [9, 10, 15, 16]. In [10] it ws sown tt t two-olorin numr is rlt to t yli romti numr. It turns out tt t two-olorin numr is lso rlt to t notion o nrt olorins introu y Boroin [1], wi ws t strtin motivtion or t rsults o tis ppr. Lt G rp, n lt L linr orrin o V (G). A vrtx x is k-rl rom y i x< L y n tr is n xy-pt P o lnt t most k su tt y< L z or ll intrior vrtis z o P.LtR L,k (y) t st o ll vrtis x tt r k-rl rom y wit rspt to t linr orr L. Tk-olorin numr o G is in s ol k (G) =1+min L mx R L,k(y), y V (G) wr t minimum is tkn ovr ll linr orrins L o V (G). I k =1,tn ol 1 (G) islsoknownstolorin numr o G sin it provis n uppr oun Riv y t itors Sptmr 25, 2007; pt or pulition (in rvis orm) July 7, 2009; pulis ltronilly DATE. ttp://www.sim.or/journls/sim/x-x/70369.tml Dprtmnt o Mtmtis n Sttistis, Arizon Stt Univrsity, Tmp, AZ 85287-1804 (kirst@su.u). Tis utor s rsr ws support in prt y NSA rnt H98230-08-1-0069. Dprtmnt o Mtmtis, Simon Frsr Univrsity, Burny, BC V5A 1S6, Cn (mor@su.). On lv rom Dprtmnt o Mtmtis, IMFM & FMF, Univrsity o Ljuljn, Ljuljn, Slovni. Tis utor s rsr ws support in prt y t ARRS (Slovni), Rsr Prorm P1 0297, y n NSERC Disovry rnt, n y t Cn Rsr Cir prorm. Univrsity o Mrior, FME, Smtnov 17, 2000 Mrior, Slovni (simon.sppn@unim.si). Tis utor s rsr ws support in prt y t ARRS (Slovni), Rsr Prorm P1 0297. Cntr or Disrt Mtmtis, Fuzou Univrsity, Fuzou, Fujin 350002, Cin (qin85@ yoo.om). Tis utor s rsr ws support in prt y NSFC unr rnts 10771035 n SX2006-42 o t oll o Fujin. Dprtmnt o Appli Mtmtis, Ntionl Sun Yt-sn Univrsity, Kosiun 80424, Tiwn, Popl s Rpuli o Cin, n Ntionl Cntr or Tortil Sins, Ntionl Tiwn Univrsity, Tipi 10617, Tiwn, Popl s Rpuli o Cin (zu@mt.nsysu.u.tw). Tis utor s rsr ws support y Tiwn rnt NSC95-2115-M-110-013-MY3. 1

2 KIERSTEAD, MOHAR, ŠPACAPAN, YANG, AND ZHU or t romti numr o G. T ir k-olorin numrs provi uppr ouns or som otr olorin prmtrs [17]. Lt k positiv intr. A rp G is k-nrt i vry surp o G s vrtx o r lss tn k. A olorin o rp su tt or vry k 1, t union o ny k olor lsss inus k-nrt surp, is nrt olorin. Not tt tis strntns t notion o yli olorins, or wi it is rquir tt vry olor lss is 1-nrt n t union o ny two olor lsss inus 2-nrt rp ( orst). T nrt romti numr o G, not s χ (G), is t lst n su tt G mits nrt olorin wit n olors. Suppos tt or vrtx v V (G) w ssin list L(v) N o missil olors wi n us to olor t vrtx v. A list olorin o G is untion : V (G) N, su tt (v) L(v) orv V (G) n(u) (v) wnvr u n v r jnt in G. I t olorin is lso nrt, w sy tt is nrt list olorin. I or ny oi o lists L(v),v V (G), su tt L(v) k, trxists list olorin o G, tn w sy tt G is k-oosl. T list romti numr (or t oi numr) og, not s (G), is t lst k, su tt G is k-oosl. Anloously w in t nrt oi numr n not it y (G). In 1976 Boroin prov [2, 3] tt vry plnr rp mits n yli 5-olorin n tus solv onjtur propos y Grünum [7]. At t sm tim, propos t ollowin onjtur. Conjtur 1 (Boroin [2, 3]). Evry plnr rp s 5-olorin su tt t union o vry k olor lsss wit 1 k 4 inus k-nrt rp. Tomssn sttl wknin o Conjtur 1 y provin tt t vrtx st o vry plnr rp n ompos into two sts tt, rsptivly, inu 2-nrt rp n 3-nrt rp [13], n tt t vrtx st o vry plnr rp n ompos into n inpnnt st n st tt inus 4-nrt rp [14]. Howvr, Conjtur 1 rmin silly untou sin tr r no tools to l wit nrt olorins. Vry rntly, t rrir ws ovrrin y Rutn [11] wo prov tt vry plnr rp mits nrt olorin usin t most 18 olors. In tis ppr w introu two irnt ppros or lin wit nrt olorins. Bot r s on t ollowin osrvtion. Osrvtion 1. Lt G rp, n lt nrt olorin o vrtxlt surp G v. I t niors o v r olor y pirwis istint olors n w olor v y olor wi is irnt rom ll o tos olors, tn t rsultin olorin o G is nrt. T osrvtion ov, wos sy proo is lt to t rr, is link twn t two-olorin numr n t nrt romti numr, sin it implis t ollowin proposition. Proposition 1. For ny rp G, (G) ol 2 (G). Proo. Suppos tt L is linr orrin o V (G), n suppos tt vrtx y s list o R L,2 (y) + 1 olors. Tn w n olor t vrtis o G, onyon, orin to t linr orrin L, sottvrtxy is olor y olor irnt rom t olors o vrtis in R L,2 (y). Tis strty urnts tt ll niors o y, wi r lry olor, v pirwis irnt olors. To s tis, lt y 1 < L y 2 ny two niors o y, wry 1,y 2 < L y. Tn y 1 R L,2 (y 2 ), so y 2 s n olor irntly rom y 1. It ollows rom Osrvtion 1 tt t otin olorin o G is nrt n n (G) ol 2 (G). In tis ppr w sll us two-olorins to t n iint oun on t nrt

THE TWO-COLORING NUMBER AND DEGENERATE COLORINGS 3 romti numr. It is prov in [9] tt plnr rps v two-olorin numr t most 10. Currntly, our min rsult ivs t st uppr oun or t two-olorin numr o vry plnr rp G, n nort oun on (G). Torm 1. I G is plnr rp, tn ol 2 (G) 9, n tror lso (G) 9. Torm 1 will prov in stion 2. In t, w will prov slitly stronr sttmnt tt tr is linr orrin L o V (G) su tt or vrtx x, R L,1 (x) 5n R L,2 (x) 8. S Torm 2. An xmpl ivn in [9] sows tt tr r plnr rps wos 2-olorin numr is 8. Tt xmpl is quit omplit. Hr w iv mu simplr on. Consir iv-onnt trinultion T o t pln in wi no two vrtis o r 5 r jnt. It is wll known n sy to s tt tr r ininitly mny su trinultions. Lt L linr orrin o vrtis o T,nltx t lst vrtx tt s lrr nior wit rspt to L. Suppos tt y 1,...,y k r t niors o x su tt x< L y i or i =1,...,k,nltN (x) tsto niors o x istint rom y 1,...,y k. By t oi o x, ll niors o y 1,...,y k r 2-rl rom x. Not tt y i stmosttwoniorsinn (x). I k = 1, it is sy to s tt N (x) n t niors o y 1 ontin t lst svn vrtis istint rom x. Tus, R L,2 (x) 7. On t otr n, i k 2, tn y 1 n y 2 v only x n possily on nior o x s ommon niors. I on o tm s r t lst 6, tn ty v t lst svn niors istint rom x, n n R L,2 (x) 7. I ty ot v r 5, tn x s r t lst 6, n it is in sy to s tt R L,2 (x) 7. Tis sows tt ol 2 (T ) 8. T ollowin rmins llnin opn prolm. Qustion 1. Is it tru tt vry plnr rp G stisis ol 2 (G) 8? u y F 1 F 2 x v Fi. 1. x n y r opposit wit rspt to = uv. T sontool win us to ontrolnrywn lin wit rps m in surs is s on t notion o ionl olorins in low. Lt G pln rp, n lt = uv n o G. Suppos tt t s F 1 n F 2 inint wit r ot trinls. I x, y r t vrtis istint rom u, v on t ounry o F 1 n F 2, rsptivly, tn w sy tt x, y r opposit wit rspt to (s Fiur 1). Vrtis x n y o G r si to opposit i ty r opposit wit rspt to som o G. Not tt or vrtx o r k, tr r t most k vrtis tt r opposit to it. Bout t l. [5] introu t notion o ionl olorins o pln rps, or wi on rquirs tt ny two jnt or opposit vrtis riv irnt olors. Ty propos t ollowin onjtur. Conjtur 2 (Bout t l. [5]). Evry pln rp s ionl 9-olorin.

4 KIERSTEAD, MOHAR, ŠPACAPAN, YANG, AND ZHU Fi. 2. A trinultion wi ns nin olors. T rp sown in Fiur 2 is n xmpl wi ns nin olors. Bout t l. prov in [5] tt 12 olors sui in nrl. Boroin [4] sow ow to sv on olor, n Snrs n Zo [12] prov tt 10 olors sui. S lso [8, Prolm 2.15]. 2. T two-olorin numr. First w n som proprtis o pln trinultions. Suppos G =(V,E) is trinultion wos vrtis r prtition into two susts U n C, wrc is n inpnnt st n vrtx in C s r 4. For vrtx x, lt U (x) t numr o niors o x in U, nlt C (x) t numr o niors o x in C. Osrvtt C (x) =4 C. x U Lt w(x) = U (x)+ C (x)/2 twit o x. Tn (1) 2 E = G (x) = G (x)+4 C = w(x)+6 C. x V x U x U Eulr s ormul implis tt E < 3 V =3 U +3 C. Tis implis tt 6 U > x U w(x), so tr is vrtx x U wit w(x) < 6 (n w(x) 5.5). ForrpG n x, y V (G) wwritx y wn x is jnt to y in G n x y otrwis. W sy tt x U is vrtx o typ (, ) i U (x) = n C (x) =. Ix y, x is o typ (5,0) or (5,1), n y is o typ (5,0), (5,1), or (5,2), tn (x, y) is ll oo pir. I xyz is trinl in G, z is o typ (5,0) or (5,1), n x, y r ot o typ (6,0), tn (x, y, z) isoo tripl. Ix, y r o typs (5,0) or (5,1) n z is o typ (6,1) n x z,y z,x y, tn(x, y, z) is lso ll oo tripl. Lmm 1. Lt G =(V,E) plnr trinultion. Suppos tt C V is n inpnnt st wr vrtx o C s r 4, ltu = V C, n suppos tt U (x) 5 or ll x U. TnG s oo pir or oo tripl. Proo. Assum tt tr is no oo pir n no oo tripl. Osrv tt U. For x U, lt(x) =w(x) 6tinitil r o x. A vrtx x is ll mjor vrtx i (x) > 0, n vrtx x wit (x) < 0 is ll minor vrtx. Eulr s ormul implis tt E =3 V 6, n totr wit (1) w onlu tt (2) (x) = w(x) 6 U =2 E 6 C 6 U = 12. x U x U

THE TWO-COLORING NUMBER AND DEGENERATE COLORINGS 5 I x is minor vrtx o typ (5,0), tn w lt mjor nior o x sn r o 1/3 tox. I x is minor vrtx o typ (5,1), tn mjor nior o x is sk to sn r o 1/6 tox. Lt us onsir t rsultin r (x) or ll vrtis o G. W sll sow tt (x) 0orx, wi will ontrit (2) n omplt t proo. Not tt vry vrtx, wi is not mjor vrtx, is o typ (5, 0), (5, 1), (5, 2), or (6, 0). I x is minor vrtx, tn its typ is (5, 0) or (5, 1). Sin tr r no oo pirs, x is not jnt to vrtx o typ (5, 0), (5, 1), or (5, 2). Aitionlly, sin tr r no oo tripls, x is not jnt to two jnt vrtis o typ (6, 0). It ollows tt vry minor vrtx s t lst tr mjor niors. Tror (x) 0 or ll minor vrtis x. Suppos now tt x is mjor vrtx. Sin tr r no oo pirs, no two minor vrtis r jnt, n n x sns r only to nonniorin vrtis. I minoru-nior y o x is jnt to C-nior o x, tny is o typ (5,1), n n x sns to y only 1/6 r. A U-nior jnt to two C-niors o x rivs 0. Tror t totl r snt out rom x is t most U (x)/2 /3. Hn (x) 0i U (x) 7. Assum now tt U (x) =6. I C (x) 2, tn it is sy to s tt t r snt out rom x is t most 2/3, n n (x) 0. So, ssum U (x) =6n C (x) =1. Ix s t lst two minor niors, tn ts two minor niors r not jnt (or otrwis w v oo pir). But tn w v oo tripl, ontrry to our ssumption. Tus x s t most on minor nior, n n t r snt out rom x is t most 1/3. Sin (x) =1/2 w onlu tt (x) 1/2 1/3 > 0. I (x) = 0, tn no r is snt out rom x, n n (x) =0. Nowvry vrtx s nonntiv r, wi is ontrition. Lt G =(V,E) rp n C V.AC-orrin o G is prtil orrin L o V (G) su tt t ollowin onitions ol: (i) T rstrition o L to C is linr orrin o C. (ii) T vrtis in V C r inomprl (tt is, nitr x< L y nor y< L x or x, y V C). (iii) y< L x or vry x C n y V C. Lt L C-orrin o G n x C. TstR L,2 (x) iststolly< L x, su tt itr y x or tr xists z C wit x z,y z, nx< L z.int lttr s w sy tt y is two-rl rom x wit rspt to L. Suppos tt v is vrtx o G n C = C {v}. IL is C -orrin o G n L is C-orrin o G su tt L n L r qul on C n v< L u or ll u C, tn lrly R L,2 (x) = R L,2(x) or ll x C. Tus, xtnin t C-orrin to C -orrin os not inrs t siz o ny st R L,2 (x). Torm 2. I G is plnr rp, tn tr is linr orrin L o V (G) su tt or vrtx x, R L,1 (x) 5 n R L,2 (x) 8. Inprtiulr,ol 2 (G) 9. Proo. W n to in linr orrin L o V (G) su tt or vrtx x, R L,1 (x) 5n R L,2 (x) 8. Lt C V n U = V C. W sy tt C-orrin L o G is vli i vrtx x C s t most our niors in U n R L,2 (x) 8. In t ollowin w sll prou squn o vli orrins o G so tt t orr st C oms lrr n lrr, n vntully prous linr orrin o V (G). At t innin C = (wi is rtinly vli orrin). Suppos w v

6 KIERSTEAD, MOHAR, ŠPACAPAN, YANG, AND ZHU vli C-orrin L o G n U. To prou lrr vli orrin w n t rp G s ollows. First w lt ll s twn vrtis o C. Ix C s t most tr niors in U, tn lt x n s twn pir o t U-niors o x. Osrv tt tis oprtion prsrvs plnrity o t rp. I x C s our niors in U, tn s in t yli orr (orin to t pln min) so tt t our niors orm 4-yl. T rsultin rp G 1 is plnr n vry vrtx o G 1 in C is o r 4 n is ontin only in trinulr s. Tror w n s only mon vrtis o U to otin pln trinultion G in wi t vrtx st F = C V (G ) is inpnnt n ll its vrtis v r 4. Sin G is trinultion n F is n inpnnt st o G, w know tt or vrtx x U, U (x) F (x). I tr is vrtx x U wit U (x) 4, tn xtn t orrin L y lttin x t nxt orr vrtx. T otin orrin L is n orrin wit x< L y or ll y C n u< L x or ll u U {x}. Lt z nior o x, wi is in C. Iz F, tn it s our niors in U n t most on o tm is not x or nior o x. Consquntly t most on nw vrtx is two-rl rom x trou z. On t otr n i z C F,tnz s t most tr niors in U n ll o tm r itr x or its niors in G. W inr tt R L,2 (x) 8 ols in t oriinl rp G. By ssumption, x s t most our niors in U, so t xtn orrin is vli. j x y i r q p x k z o y j i n m n k s m l l Fi. 3. A s o n xtnsion o orrin, wr ot x n y r o typ (5,1). Dionl vrtis r init y rokn lins. Vrtis o F r ull; vrtis o U r ollow. Suppos now tt U (x) 5inG or x U. W will pply Lmm 1 or G n t st F plyin t rol o C. T lmm implis tt tr is itr oo pir (x, y) or oo tripl (x, y, z). I tr is oo pir (x, y), tn w xtn t orrin y rqustin tt or ny u U {x, y} n v C w v u< L x< L y< L v. W lim tt t otin orrin L is vli in G or t prorr st C 1 = C {x, y}. W trt t s wn x n y r o typ (5,1) in til, n lv ll otr ss to t rr. Lt t nior o x in F,nlt t nior o y in F (possily = ). Not tt n ror4ing n tt o tm s t most on nior tt is two-rl rom x (rsp., y) n is not nior o x (rsp., y). An xmpl o tis s is sown in Fiur 3. Not tt t typ (5,1) n t rwin in Fiur 3 r wit rspt to G. T uniqu niors o x n y in F r osn in Fiur 3 s vrtis n, rsptivly, ut ty n ny otr niors, inluin possily in on o t ommon niors o x n y. Sin t xtnsion

THE TWO-COLORING NUMBER AND DEGENERATE COLORINGS 7 L o L is in so tt x< L y, w s tt t numr o niors u o x wit u< L x is our (in Fiur 3 ts r,,, n). Morovr, t numr o vrtis tt r two-rl in G rom x (n r irnt rom t U-niors o x in G ) is t most tr, on trou n two mor trou y (ts r j,, n). On t otr n, y s iv niors u wit u< L y (nmly, x,,, n) n on itionl two-rl vrtx trou tt is not its nior in G,ttisi. W prov tt R L,2(x), R L,2(y) 7nox, y s t most our niors in U {x, y}. Tror t xtn orrin is vli. Otr oniurtions, wn (x, y) is oo pir, r trt similrly. I (x, y, z) is oo tripl su tt x n y r o typs (5,0) or (5,1) n z is o typ (6,1),tnxtntC-orrin L to L y inin positions or x, y, nz s ollows: or ny u U {x, y, z} n v C, wstu< L x< L z< L y< L v. I (x, y, z) is oo tripl su tt x n y r o typ (6,0) n z is o typ (5,0) or (5,1),tnxtntC-orrin L to L y sttin u< L x< L y< L z< L v or u U {x, y, z} n v C. As or, it is sy to vriy tt R L,2(x), R L,2(y), R L,2(z) 8nox, y, z s t most our niors in U {x, y, z}. W onlu tt L is vli orrin. Not lso tt t nwly in prtil orr L stisis R L,1(x) 5 or ll x. On o t ss is sown in Fiur 3 or t onvnin o t rr. A slit n in t ov proo yils n pplition to nrt ionl list olorins usin t most 12 olors t vrtx. Corollry 1. Evry pln rp s nrt ionl list olorin rom lists o siz t lst 12. Proo. Lt L C-orrin, n lt RL,2 (y) t st o ll vrtis tt r two-rl rom y, totr wit ll vrtis x, su tt x< L y n x is opposit y. For t purpos o tis proo w sy tt L is vli orrin i or vrtx x C, RL,2 (x) 11, n x s t most our niors in U. Wprovtxistn o linr orr L so tt RL,2 (y) 11 or vrtx y V (G). W pro nloously s in t proo o Torm 1, tt is, w onstrut squn o vli orrins until w vntully t linr orrin o V (G). So suppos tt w r ivn vli C-orrin L, wrc V (G). Tn n t rp G s in t proo o Torm 1 to otin pln rp G. It is sy to s tt t onstrution o G ws on so tt i x, y U n x is opposit y in G, tnx is itr jnt to or opposit y in G. Ain w rriv t w ss, wr w v to trmin ow to xtn t orrin L. I tr is vrtx x wit U (x) 4inG, tn xtn t orrin L to L so tt t nxt orr vrtx is x n osrv tt ny vrtx tt is two-rl rom x n is not nior o x is lso opposit to x. It ollows tt RL,2(y) 8 ols in G. Otrwis G s itr oo pir or oo tripl. I (x, y) is oo pir, tn lt x< L y. I x, y r o typ (5,1) (s Fiur 3), tn osrv tt x s our jnt vrtis smllr tn x (ts r,,, ), two vrtis wi r tworl n ionl t t sm tim (ts r j n ), two ionl vrtis wi r not two-rl (ts r m n n) n on otr two-rl vrtx (). Tis is nin ltotr, n n RL,2(x) 9. It is sy to k tt RL,2(y) ={, x,,,, i, k, l, }, so its siz is 9. Otr ss o oo pirs r trt similrly ut RL,2 (x) n R L,2(y) r lwys t most 10. I (x, y, z) is oo tripl su tt x n y r o typs (5,0) or (5,1) n z is o typ (6, 1), w xtn L to L so tt z < L x< L y (s Fiur 3). Tn

8 KIERSTEAD, MOHAR, ŠPACAPAN, YANG, AND ZHU w v RL,2 (z) ={, k,,, l, s,, j,, }. Similrly R L,2(x) ={k, z,,,,, q, r, p} n RL,2(y) ={,, z, k, j, m, n, k, o}, n n in ny s t siz is lss tn 10. I (x, y, z) is oo tripl wr z is o typ (5,0) or (5,1) n x, y r o typ (6,0), tn lt x < L y < L z, n it is sy to k tt RL,2 (x), R L,2 (y), RL,2(z) 11. 3. Dnrt ionl 11-olorins. T ollowin torm is n improvmnt o Corollry 1 y on olor or t nonlist vrsion o nrt n ionl olorins. W v i to inlu it spit t t tt it is slitly lonr sin t mtos us in t proo o tis rsult r irnt rom t mtos o t prvious stion. Torm 3. Evry pln rp s nrt ionl olorin wit 11 olors. Proo. Suppos tt t torm is not tru. Lt G minimum ountrxmpl (wit rspt to t numr o vrtis). W my ssum tt G is trinultion o t pln. Clim 0. G s no vrtis o r 3. Proo. Suppos, on t ontrry, tt v V (G) is vrtx o r 3. Sin G is minimum ountrxmpl, tr is nrt olorin o G v, su tt no two opposit vrtis r olor wit t sm olor. A sir olorin o G is otin rom t olorin o G v, y olorin t vrtx v wit olor irnt rom t olors o its niors n its opposit vrtis. Not tt tis is lwys possil, sin tr r t most six niors or opposit vrtis o v; tror in t st o 11 olors tr r 5 possil olors or v. Tis is ontrition to t oi o G. Clim 1. G s no vrtis o r 4. Proo. Suppos, on t ontrry, tt v is vrtx o r 4. S Fiur 4. Witout loss o nrlity ssum tt is not n in G. Dlt t vrtx v, t, n ll t otin rp G. By t minimlity o G, tris nrt olorin o G su tt n r olor wit istint olors (sin ty r opposit). Tis olorin inus nrt olorin o G v. InG tr r t most our vrtis opposit to v n tr r our vrtis jnt to v. W olor t vrtx v wit olor irnt rom olors o t vrtis jnt or opposit to v. Sin,,, n r olor y pirwis istint olors, w inr rom Osrvtion 1 tt t otin olorin is nrt. Morovr, ny two opposit vrtis r olor wit istint olors. v Fi. 4. A vrtx o r 4. Lt us now introu som nottion. I G s sprtin 3-yl, tn lt C sprtin 3-yl wit t lst numr o vrtis in its intrior. Otrwis lt C t 3-yl on t ounry o t ininit. Dnot y C t rp inu y vrtis o C n tos in t intrior o C. I C ontins 4-yl or 5-yl wit t lst two vrtis in its intrior, tn lt D t on wit t lst numr o vrtis in its intrior; otrwis, lt D = C. Lt D t rp inu y t

THE TWO-COLORING NUMBER AND DEGENERATE COLORINGS 9 vrtis o D n tos in t intrior o D. W sll not y int(d) tsto intrnl vrtis o D. Clim 2. In C, no two intrnl vrtis o r 5 r jnt. Proo. Suppos, on t ontrry, tt u, v r jnt intrnl vrtis o r 5; s Fiur 5 or nottion. Sin C s no sprtin 3-yls,,, n r not s o G. Dlt t vrtis u n v, t s,, n, nllt otin rp G (s Fiur 5). u v Fi. 5. Two intrnl jnt vrtis o r 5. Osrv tt,, n, r two pirs o opposit vrtis in G ; tror tr xists nrt olorin o G su tt t vrtx sts {,,, } n {,,, } r olor y pirwis istint olors. Sin ot u n v r vrtis o r 5, tr r t most nin (iv on opposit n our on jnt vrtis) olors proiit or u n v. Tror, olorin o G n otin rom t olorin o G y olorin u n v y istint vill olors. Osrv tt t otin olorin is nrt (y Osrvtion 1) n tt ny two opposit vrtis r olor y istint olors. v u z x, i i Fi. 6. u, v, z, nx r vrtis o rs 5, 6, 6, n6, rsptivly. Clim 3. An intrnl vrtx u o D o r 5 nnot jnt to tr intrnl vrtis v, x, z o D o r 6, wrz is jnt to x n v. Proo. Suppos tt u, v, z, nx r intrnl vrtis o D ontritin t lim (s Fiur 6). W lim tt vrtis i n r nitr jnt nor v ommon nior. Assum (or ontrition) tt ty r jnt n tt t i is m s sown in Fiur 6. Consir t 4-yl E = iuv n osrv tt it ontins t lst two vrtis in t intrior. By t minimlity o D, t 4-yl E is not ontin in D. So ssum tt t is vrtx o D ontin in t intrior o E. I t is jnt to ot n i, tntvuit is 5-yl ontritin t minimlity o D. Otrwis, tr r two vrtis o D ontin in t intrior o E n vrtx t V (D) in t xtrior o E; intisst is jnt to n

10 KIERSTEAD, MOHAR, ŠPACAPAN, YANG, AND ZHU i. Ain w onlu tt t vuit is 5-yl, ontritin t minimlity o D. Anloous rumnts prov tt n i nnot v ommon nior. Not lso tt t vrtis,,...,i sown in Fiur 6 r pirwis istint, sin otrwis D woul ontin sprtin 3-yl or 4-yl (irnt rom D). Morovr, t sm rumnts sow tt non o,,,, n is n o G. Lt us lt vrtis u, v, z, nx n intiy i n. Furtr, t s,, n s sown in Fiur 6, n ll t otin rp G. As prov ov, G is rp witout loops or multipl s. By t minimlity o G, tr is nrt olorin o G, su tt o t vrtx sts {i,,, }, {,, }, {,,, }, n{, i} is olor y pirwis istint olors (not tt i,, n, r pirs o opposit vrtis in G ). Tis olorin inus olorin o ( surp o) G, wri n riv t sm olor. Sin n i r t istn t lst 3inG, tis olorin is nrt n s opposit vrtis olor wit istint olors. W now olor t vrtis x, z, v, nu in tis orr. Sin x is olor or u, v, nz n t olor o quls t olor o i, w in tt tr r t most nin proiit olors or x; ts r t olors us or t vrtis,,,,, i, n tr itionl opposit vrtis istint rom, v, n. W olor t vrtx x wit on o t two rminin olors not proiit or x. Nxt w olor z, wi s 10 proiit olors; ts r t olors us or t vrtis,,, x,,,,, nt two opposit vrtis wit rspt to s n. Notttwvrqust tt t olor o z irnt rom in orr to l to pply Osrvtion 1 wn olorin v. So tr is olor not proiit or z, n w us tis olor to olor it. Sin i n v t sm olor, t olor o x is istint rom t olor o. Hn, t olors on t niors o z r ll istint. Tror t olorin is still nrt. Nxt w olor v, wi s t most 10 proiit olors: it s six niors n six opposit vrtis, ut u is not yt olor, n, i v t sm olor. Finlly, w olor u, wi s lso t most nin proiit olors (t olors o iv opposit n iv jnt vrtis, wr two o tm oini, nmly t olors o n i). By pplyin Osrvtion 1 t stp, w onlu tt t rsultin olorin o G is nrt. It ws onstrut in su wy tt ll opposit vrtis v irnt olors. Tis ontrition to nonolorility o G omplts t proo o Clim 3. Clim 4. An intrnl vrtx u o D o r 7 nnot jnt to tr intrnl vrtis v, x, z o D o r 5 n two vrtis y n t o r 6, su tt y is jnt to v n x n t is jnt to x n z. Proo. Suppos tt u, v, x, z, y, t r vrtis o D ontritin t lim. S Fiur 7 or itionl nottion. Sin u, v, z, nx r intrnl vrtis o D, w in tt n r nitr jnt nor v ommon nior (sin otrwis on woul in 4- or 5-yl wit wr vrtis in t intrior wn ompr to D). Lt us rmov vrtis t, u, v, x, y, z. Tn intiy vrtis n, s i, i,, n (or ), n ll t otin rp G. T nrt olorin o G inus nrt olorin o surp o G, wr n riv t sm olor. W will olor t vrtis u, y, t, x, v, z in tis orr. Osrv tt or t vrtx u, tr r t most 10 = 14 + 2 5 1 proiit olors: 14 olors or jnt n opposit vrtis, +2 olors or j n (wi w wnt to olor irnt rom u to us Osrvtion 1 wn ltr olorin y n t), 5 or not yt olor jnt vrtis, n 1 us t olor o is qul to t olor o. Similrly on n

THE TWO-COLORING NUMBER AND DEGENERATE COLORINGS 11 i v y u t z i j x j Fi. 7. u, v, z, nx r vrtis o rs 7, 5, 5, n5, rsptivly. ru tt or y tr r t most it rstritions, n or t, v, x, z, tr r t most 10 proiit olors. Sin w olor wit 11 olors n or vrtx u, y, t, x, v, z tr r wr tn 11 proiit olors, w n xtn t prtil olorin o G inu y t olorin o G to nrt ionl olorin o G. Clim 4 implis tt, i u is vrtx o r 7 wit tr (nonjnt) niors v, x, z o r 5 n istriut s sown in Fiur 7, tn itr y or t s r t lst 7. W omplt t proo o t torm y pplyin t isrin mto in D. W in t r (v) ovrtxv V ( D) s { D(v) 6 i v int(d), (3) (v) = D(v) 3 i v V (D). It ollows rom t Eulr ormul tt ( D(v) 6) = 6 2 D, n tus (4) v V ( D) (v) = v V (D) v V ( D) ( D(v) 3) + v int(d) ( D(v) 6) = D 6. W now prorm t isrin prour s ollows: (i) Evry vrtx o D ivs 1 to nior o r 5 in int(d). (ii) Evry vrtx in int(d) o r t lst 8 ivs 1/2 to nior o r 5inint(D). (iii) Evry vrtx in int(d) o r 7 wit t most two niors o r 5 in int(d) ivs 1/2 to nior o r 5 in int(d). (iv) A vrtx u int(d) o r 7 wit tr niors o r 5 in int(d) s its nioroo s sown in Fiur 7, xpt tt y n t r not nssrily o r 6. () I (y) 7 n (t) 7, tn u ivs 1/2 to x n 1/4 to v n z. () I (y) 7 n (t) =6,tnu ivs 1/2 to z n 1/4 to v n x. () I (y) = 6 n (t) 7, tn u ivs 1/2 to v n 1/4 to z n x. Not tt vrtx o D nnot v r 3in D, sin tis woul imply tt D is not 3-, 4-, or 5-yl wit minimum numr o vrtis in its intrior.

12 KIERSTEAD, MOHAR, ŠPACAPAN, YANG, AND ZHU Morovr, i v is vrtx o D n s α>1 niors o r 5 in int(d), tn t r o v in D is t lst 2α + 1 y Clim 2. Tis implis tt vry vrtx o D rtins (tr isrin) t lst 0 o its r. Also, y Clim 2, vrtx o r α 7 s t most α 2 niors o r 5. So w inr rom t isrin ruls tt vry vrtx o int(d) or 7 s nonntiv r tr t isrin prour. Sin intrnl vrtis o r 6 kp tir r t zro, n tr r no vrtis o r lss tn 5; t only nits or vin ntiv r r intrnl vrtis o r 5. Osrv tt vry vrtx v int(d) o r 5 is itr jnt to vrtx o D or s, y Clims 0, 1, 2, n 3, t lst two niors o r 7. I v s our or iv niors o r 7, tn, tr isrin, its r will t lst 0, sin vry vrtx o r 7 ivs t lst 1/4 to o its niors o r 5. I v s tr niors o r 7 n on o tm is not o r 7 or os not v tr niors o r 5, tn tis nior will iv 1/2 to v n tus v will v r o t lst 0. Otrwis v s tr niors o r 7, n o tm s tr niors o r 5. It ollows rom rul (iv) tt on o t niors o v will iv 1/2 to v, so t inl r o v will nonntiv. In t rminin s, wr v s xtly two niors o r 7, w s tt ts two niors r not jnt y Clim 3. Tror t isrin rul (iv) implis tt ot o tm iv 1/2 to v. Tus, t inl r t v is nonntiv. Tis isrin pross provs tt t lt-n si o (4) is nonntiv, wil t rit-n si is ntiv, ontrition. Tis omplts t proo o Torm 3. Finlly w turn to t proo o t ionl list olorin rsult. Osrv tt intiition o vrtis nnot on wit list olorins, so t proos o Clims 3 n 4 rom t ov proo nnot xtn to list olorins. W nxt iv irnt proo o Corollry 1. T proo is similr to t proo o Torm 3. W strt y ssumin t ontrry, n lt G minimum ountrxmpl. Din C, C,D,n D t sm wy s in t proo o Torm 3. T proos o Clims 0, 1, n 2 rom t proo o Torm 3 lso ol or list olorins. Clim 5. An intrnl vrtx o D o r 5 nnot jnt to two jnt intrnl vrtis o D o r 6. Proo. Suppos, on t ontrry, tt u, v, z r vrtis o rs 5, 6, n 6, rsptivly. S Fiur 8 or urtr nottion. W o t rution s ollows. v u z Fi. 8. u, v, nz r vrtis o rs 5, 6, n6, rsptivly. Dlt t vrtis u, v, nz n s,,, n. Not tt ts r not s o G, sin D s no sprtin 3-yls. Lt us ll t otin rp

THE TWO-COLORING NUMBER AND DEGENERATE COLORINGS 13 G. T lists or vrtis o G r inrit G. By t minimlity o G, tris nrt list olorin o G, su tt,,, n,,,, rsptivly, r olor y pirwis istint olors (sin,, n, r pirs o opposit vrtis in G ). Tis olorin inus olorin o surp o G. W nxt olor v, z, nu (in tis orr). Osrv tt v n z v t most 11 proiit olors (6 opposit vrtis n 4, rsptivly 5, niors, n t olor o is proiit or v); t vrtx u is o r 5, n tus it s t most 10 proiit olors. Tus, or ll tr rminin vrtis tr is r olor so tt w n xtn t list olorin o G to nrt list olorin o G; tis is ontrition provin t lim. Finlly, t ollowin isrin prour ls to ontrition. (i) Evry vrtx o D ivs 1 to nior o r 5 in int(d). (ii) Evry vrtx o r t lst 7 o int(d) ivs 1/3 to nior o r 5inint(D). Sin vry vrtx o r 5 o int(d) s (y Clims 0, 1, 2, n 5) t lst tr niors o r 7 or s nior in D, n vry vrtx o r α 7o int(d) s t most α 2 niors o r 5 in int(d), w onlu tt t lt-n si o (4) is nonntiv, wil t rit-n si is ntiv, ontrition. REFERENCES [1] O. V. Boroin, On omposition o rps into nrt surps, Diskrtny Anliz., 28 (1976), pp. 3 12 (in Russin). [2] O. V. Boroin, A proo o Grünum s onjtur on t yli 5-olorility o plnr rps, Sovit Mt. Dokl., 17 (1976), pp. 1499 1502. [3] O. V. Boroin, On yli olorins o plnr rps, Disrt Mt., 25 (1979), pp. 211 236. [4] O. V. Boroin, Dionl 11-olorin o pln trinultions, J. Grp Tory, 14 (1990), pp. 701 704. [5] A. Bout, J.-L. Fouqut, J.-L. Jolivt, n M. Riviér, On spil olourin o ui rps, Ars Comin., 24 (1987), pp. 67 76. [6] G. Cn n R. H. Slp, Grps wit linrly oun Rmsy numrs, J. Comin. Tory Sr. B, 57 (1993), pp. 138 149. [7] B. Grünum, Ayli olorins o plnr rps, Isrl J. Mt., 14 (1973), pp. 390 408. [8] T. R. Jnsn n B. Tot, Grp Colorin Prolms, Jon Wily & Sons, Nw York, 1995. [9] H. A. Kirst n W. T. Trottr, Plnr rp olorin wit n unooprtiv prtnr, J. Grp Tory, 18 (1994), pp. 569 584. [10] H. A. Kirst n D. Yn, Orrins on rps n m olorin numr, Orr, 20 (2003), pp. 255 264. [11] D. Rutn, A onjtur o Boroin n olorin o Grünum, Fit Crow Conrn on Grp Tory, Ustron 06, Eltron. Nots Disrt Mt., 24 (2006), pp. 187 194. [12] D. P. Snrs n Y. Zo, On ionlly 10-olorin pln trinultions, J. Grp Tory, 20 (1995), pp. 77 85. [13] C. Tomssn, Domposin plnr rp into nrt rps, J. Comin. Tory Sr. B, 65 (1995), pp. 305 314. [14] C. Tomssn, Domposin plnr rp into n inpnnt st n 3-nrt rp, J. Comin. Tory Sr. B, 83 (2001), pp. 262 271. [15] X. Zu, T m olorin numr o plnr rps, J. Comin. Tory Sr. B, 75 (1999), pp. 245 258. [16] X. Zu, Rin tivtion strty or t mrkin m, J. Comin. Tory Sr. B, 98 (2008), pp. 1 18. [17] X. Zu, Colourin rps wit oun nrliz olourin numr, Disrt Mt., to ppr.