1: %Format: plain

2: \input amstex

3: \documentstyle{amsppt}

4: 

5: \input label.def

6: \input degt.def

7: %\input debug.def

8: %\PrintLabels

9: 

10: {\catcode`\@11

11: \gdef\proclaimfont@{\sl}}

12: 

13: \newtheorem{Remark}\Remark\thm\endAmSdef

14: \newtheorem{Conjecture}\conjecture\thm\endproclaim

15: 

16: \input epsf

17: \def\picture#1{\epsffile{#1-bb.eps}}

18: \def\plot#1{\centerline{\epsfxsize.8\hsize\picture{plot-#1}}}

19: 

20: \def\dash{\item"\hfill--\hfill"}

21: \def\Dashes{\widestnumber\item{--}\roster}

22: \def\endDashes{\endroster}

23: 

24: \def\CR{\\\allowdisplaybreak}

25: 

26: \def\ie{\emph{i.e.}}

27: \def\eg{\emph{e.g.}}

28: \def\cf{\emph{cf}}

29: \def\via{\emph{via}}

30: \def\etc{\emph{etc}}

31: 

32: \let\Ga\alpha

33: \let\Gb\beta

34: \let\Gg\gamma

35: \let\Gd\delta

36: \let\Gs\sigma

37: \let\Gf\varphi

38: \let\Gr\rho

39: \def\bGa{\bar\Ga}

40: \def\bGb{\bar\Gb}

41: \def\bGg{\bar\Gg}

42: 

43: \def\1{^{-1}}

44: \def\2{^{-2}}

45: 

46: \def\xt{x_t}

47: \def\yt{y_t}

48: 

49: \def\bpi{\bar\pi_1}

50: 

51: \loadbold

52: \def\bA{\bold A}

53: \def\bD{\bold D}

54: \def\bE{\bold E}

55: \def\bJ{\bold J}

56: \def\bX{\bold X}

57: \def\bY{\bold Y}

58: \def\bW{\bold W}

59: 

60: \let\splus\oplus

61: \let\bigsplus\bigoplus

62: \let\onto\twoheadrightarrow

63: 

64: \def\Conv{\operatorname{conv}}

65: \def\Aut{\operatorname{Aut}}

66: 

67: \def\CG#1{\Z_{#1}}     % the cyclic group

68: \def\BG#1{\Bbb B_{#1}} % the braid group

69: \def\SG#1{\Bbb S_{#1}} % the symmetric group

70: \def\AG#1{\Bbb A_{#1}} % the alternating group

71: \def\DG#1{\Bbb D_{#1}} % the dihedral group

72: 

73: \def\Cp#1{\Bbb P^{#1}}

74: \def\Rp#1{\Bbb P^{#1}_\R}

75: \def\B{\bar B}

76: \def\L{\bar L}

77: \def\CM{\Cal M}

78: \def\tCM{\tilde\CM}

79: 

80: \def\MB{D} % Milnor ball

81: \def\disk{\Delta}

82: 

83: \def\PSL{\operatorname{\text{\sl PSL}}}

84: \def\SL{\operatorname{\text{\sl SL}}}

85: \def\gcd{\operatorname{g.c.d.}}

86: \def\<#1>{\langle#1\rangle}

87: \def\ls|#1|{\mathopen|#1\mathclose|}

88: 

89: {\let\1\gdef\let\+\relax

90: %\gdef\setcat{\def\2##1{\catcode`##1\active}\2\a \2\d \2\e \catcode`\+\active}

91: \gdef\setcat{\catcode`\a\active \catcode`\d\active \catcode`\e\active

92: \catcode`\+\active}

93: \setcat

94: \1a[#1]{\bA_{#1}} \1d[#1]{\bD_{#1}} \1e[#1]{\bE_{#1}}

95: }

96: {\catcode`\+\active\gdef+{\splus}}

97: \def\getline#1,#2[#3],#4[#5],#6,#7[#8]{#2&#1\cr}

98: {\obeylines%

99: \gdef\TAB#1{\vtop\bgroup\openup1pt\obeylines\let

100: \getline\setcat\expandafter\halign\expandafter\bgroup#1\cr}}

101: \def\ENDTAB{\crcr\egroup\egroup}

102: \def\Torus{\llap{$##$\ }&$##$\hss}

103: \def\NonTorus{$##$\hss&$##$\hss}

104: 

105: \def\Tab{\vtop\bgroup\openup1pt\halign\bgroup$##$\hss\cr}

106: 

107: \topmatter

108: 

109: \author

110: Alex Degtyarev

111: \endauthor

112: 

113: \title

114: Fundamental groups of symmetric sextics

115: \endtitle

116: 

117: \address

118: Department of Mathematics,

119: Bilkent University,

120: 06800 Ankara, Turkey

121: \endaddress

122: 

123: \email

124: degt\@fen.bilkent.edu.tr

125: \endemail

126: 

127: \abstract

128: We study the moduli spaces and compute the fundamental groups of

129: plane sextics of torus type with at least two type~$\bold{E}_6$

130: singular points. As a simple

131: application, we compute the fundamental

132: groups of $125$ other sextics, most of which are new.

133: \endabstract

134: 

135: \keywords

136: Plane sextic, torus type, fundamental group, symmetry, trigonal curve

137: \endkeywords

138: 

139: \subjclassyear{2000}

140: \subjclass

141: Primary: 14H30; % curves/Coverings, fundamental group

142: Secondary: 14H45 % curves/Special curves and curves of low genus

143: \endsubjclass

144: 

145: \endtopmatter

146: 

147: \document

148: 

149: \section{Introduction}

150: 

151: 

152: \subsection{Principal results}

153: Recall that a plane sextic~$B$ is said to be of \emph{torus type}

154: if its equation can be represented in the form $p^3+q^2=0$, where

155: $p$ and~$q$ are certain homogeneous polynomials of degree~$2$

156: and~$3$, respectively. Alternatively, $B\subset\Cp2$ is of torus

157: type if and only if it is the ramification locus of a projection

158: to~$\Cp2$ of a cubic surface in~$\Cp3$. A representation of the

159: equation in the form $p^3+q^2=0$ (up to the obvious equivalence)

160: is called a \emph{torus structure} of~$B$. A singular point~$P$

161: of~$B$ is called \emph{inner} (\emph{outer}) with respect to a

162: torus structure~$(p,q)$ if $P$ does (respectively, does not)

163: belong to the intersection of the conic $\{p=0\}$ and the cubic

164: $\{q=0\}$. The sextic~$B$ is called \emph{tame} if all its

165: singular points are inner. Note that, according

166: to~\cite{degt.Oka}, each sextic~$B$ considered in this paper has a

167: unique torus structure; hence, we can speak about inner and outer

168: singular points of~$B$.

169: For the reader's convenience, when

170: listing the set of singularities of a sextic of torus type,

171: we indicate the inner singularities by

172: enclosing them in parentheses.

173: 

174: Apparently, it was O.~Zariski~\cite{Zariski} who first understood

175: the importance of sextics of torus type. Since then, they have

176: been a subject of intensive study. For details and further

177: information,

178: we refer to

179: M.~Oka, D.~T.~Pho~\cite{OkaPho.moduli},

180: \cite{OkaPho} (topology, sets of singularities, moduli,

181: fundamental groups), H.~Tokunaga~\cite{Tokunaga} (algebro-geometric

182: approach), and A.~Degtyarev~\cite{degt.Oka}.

183: %(cohomological approach, torus structures).

184: 

185: In recent paper~\cite{degt.8a2}, we described the moduli spaces

186: and calculated the fundamental groups of all sextics of torus type

187: of weight~$8$ and~$9$ (in a sense, those with the largest

188: fundamental groups). The approach used in~\cite{degt.8a2},

189: reducing sextics to maximal

190: trigonal curves, was also helpful in the study

191: of some other sextics with nonabelian groups

192: (see~\cite{degt.Oka3}), and then, in~\cite{symmetric}, we

193: classified all irreducible sextics for which this approach should

194: work. The purpose of this paper is to treat one of the classes

195: that appeared in~\cite{symmetric}: sextics with at least two

196: type~$\bE_6$ singular points; they are reduced to trigonal curves

197: with the set of singularities $\bE_6\splus\bA_2$.

198: Our principal results are

199: Theorems~\ref{th.moduli} and~\ref{th.group} below.

200: 

201: \midinsert

202: \table

203: Sextic with two type~$\bE_6$ singular points

204: \endtable\label{tab.list}

205: \hbox to\hsize{\hss\def\1{\llap{$^*$\,}}

206: \Tab

207: \1 (3\bE_6)\splus\bA_1\cr

208:    (3\bE_6)\cr

209: \1 (2\bE_6\splus\bA_5)\splus\bA_2\cr

210:    (2\bE_6\splus\bA_5)\splus\bA_1\cr

211:    (2\bE_6\splus\bA_5)

212: \ENDTAB\hss\Tab

213: \1 (2\bE_6\splus2\bA_2)\splus\bA_3\cr

214: \1 (2\bE_6\splus2\bA_2)\splus\bA_2\cr

215: \1 (2\bE_6\splus2\bA_2)\splus2\bA_1\cr

216:    (2\bE_6\splus2\bA_2)\splus\bA_1\cr

217:    (2\bE_6\splus2\bA_2)

218: \ENDTAB

219: \hss}

220: %% 10 sets, 11 families

221: \endinsert

222: 

223: \theorem\label{th.moduli}

224: Any sextic of torus type with at least two

225: type~$\bE_6$ singular points has one of the sets of singularities

226: listed in Table~\ref{tab.list}. With the exception of

227: $(2\bE_6\splus\bA_5)\splus\bA_2$, the moduli space of sextics of

228: torus type realizing each set of singularities in the table is

229: rational \rom(in particular, it is nonempty and

230: connected\rom)\rom; the moduli space of sextics with the set of

231: singularities $(2\bE_6\splus\bA_5)\splus\bA_2$ consists of two

232: isolated points, both of torus type.

233: \endtheorem

234: 

235: Note

236: that we do not assume \emph{a priori} that the curves are

237: irreducible or have simple singularities only.

238: Both assertions hold automatically for any sextic with

239: at least two type~$\bE_6$ singular points, see the beginning of

240: Section~\ref{proof.moduli}.

241: 

242: Theorem~\ref{th.moduli} is proved in Section~\ref{proof.moduli}.

243: The two classes of sextics realizing the

244: set of singularities $(2\bE_6\splus\bA_5)\splus\bA_2$ were first

245: discovered in Oka, Pho~\cite{OkaPho.moduli}.

246: The sets of singularities that can be

247: realized by sextics of torus type are

248: also listed in~\cite{OkaPho.moduli}.

249: Note

250: that the list

251: %of singularities

252: given by Table~\ref{tab.list}

253: can also be obtained from the results of J.-G.~Yang~\cite{Yang},

254: using the characterization of irreducible

255: sextics of torus type found

256: in~\cite{degt.Oka}. The deformation classification can be obtained

257: using~\cite{JAG}.

258: 

259: \Remark

260: A simple calculation using~\cite{JAG} or~\cite{Yang}

261: and the characterization of irreducible

262: sextics of torus type found

263: in~\cite{degt.Oka}

264: shows that the sets of

265: singularities marked with a $^*$ in Table~\ref{tab.list} are realized by

266: sextics of torus type only. Each of the

267: remaining

268: five sets of

269: singularities is also realized by a single deformation family of

270: sextics not of torus type, see A.~\"Ozg\"uner~\cite{Aysegul} for

271: details. Furthermore, Table~\ref{tab.list} lists all sets of

272: singularities

273: of plane sextics, both of and not of torus type,

274: containing at least two

275: type~$\bE_6$ points.

276: \endRemark

277: 

278: \theorem\label{th.group}

279: Let $B$ be a sextic of torus type whose set of

280: singularities~$\Sigma$ is one of those listed in

281: Table~\ref{tab.list}. Then the fundamental group

282: $\pi_1:=\pi_1(\Cp2\sminus B)$ is as follows\rom:

283: \roster

284: \item\local{G3.group}

285: if $\Sigma=(2\bE_6\splus2\bA_2)\splus\bA_3$, then $\pi_1$ is the

286: group~$G_3$ given by~\eqref{eq.G3}\rom;

287: \item\local{G0.group}

288: if $\Sigma=(3\bE_6)\splus\bA_1$ or $(2\bE_6\splus2\bA_2)\splus2\bA_1$,

289: then $\pi_1=G_0:=\BG4/\Gs_2\Gs_1^2\Gs_2\Gs_3^2$\rom;

290: \item\local{two.groups}

291: if $\Sigma=(2\bE_6\splus\bA_5)\splus\bA_2$, then, depending on the

292: family, $\pi_1$ is one of

293: the groups~$G_2'$, $G_2''$ given by~\eqref{eq.G2.1}

294: and~\eqref{eq.G2.2}, respectively\rom;

295: \item\local{braid.group}

296: otherwise, $\pi_1=\BG3/(\Gs_1\Gs_2)^3$.

297: \endroster

298: %\rom(Here, $\BG{n}$ is the braid group on $n$ strings, and

299: %$\Gs_1,\ldots,\Gs_{n-1}$ is a canonical basis for~$\BG{n}$.\rom)

300: \rom(Here, $\{\Gs_1,\ldots,\Gs_{n-1}\}$ is a canonical basis for

301: the braid group~$\BG{n}$ on $n$ strings.\rom)

302: \endtheorem

303: 

304: The fundamental groups are calculated in \S\ref{S.group}. An

305: alternative presentation of the groups~$G_2'$, $G_2''$

306: mentioned in~\iref{th.group}{two.groups} is found in C.~Eyral,

307: M.~Oka~\cite{EyralOka}, where it is conjectured that

308: the two groups are not isomorphic. We suggest to attack this

309: problem studying the relation between~$G_2'$, $G_2''$ and the local

310: fundamental group at the type~$\bA_5$ singular point, \cf.

311: Proposition~\ref{a5->>pi} and Conjecture~\ref{a5not->>pi}. The

312: group of a sextic of torus type with the set of singularities

313: $(2\bE_6\splus\bA_5)\splus\bA_1$, see~\iref{th.group}{braid.group},

314: is also found in~\cite{EyralOka}; the group of a sextic with the

315: set of singularities $(3\bE_6)\splus\bA_1$,

316: see~\iref{th.group}{G0.group}, as well as the groups of the three

317: tame sextics listed in Table~\ref{tab.list} (the sets of

318: singularities $(3\bE_6)$, $(2\bE_6\splus\bA_5)$, and

319: $(2\bE_6\splus2\bA_2)$)

320: are found in Oka, Pho~\cite{OkaPho}.

321: 

322: With the possible exception of $G_2'$, $G_2''$, all groups listed in

323: Theorem~\ref{th.group} are `geometrically' distinct in the

324: sense of the following theorem.

325: 

326: \theorem\label{th.proper}

327: All epimorphisms

328: $$

329: G_3\onto G_0\onto\BG3/(\Gs_1\Gs_2)^3,\qquad

330: G_2',G_2''\onto\BG3/(\Gs_1\Gs_2)^3

331: $$

332: induced by the respective

333: perturbations of the curves \rom(\cf. Zariski~\cite{Zariski}\rom)

334: are proper, \ie, they are not isomorphisms.

335: \endtheorem

336: 

337: This theorem is proved in Section~\ref{proof.proper}. Some of the

338: statements follow from the previous results by Eyral,

339: Oka~\cite{EyralOka} and Oka, Pho~\cite{OkaPho}.

340: 

341: As a further application of Theorem~\ref{th.group},

342: we use the presentations obtained and the

343: results of~\cite{degt.8a2} to compute the fundamental groups of

344: eight sextics of torus type and $117$ sextics not of torus type

345: that are not covered by M.~V.~Nori's theorem~\cite{Nori}, see

346: Theorems~\ref{th.nontorus} and~\ref{th.torus}. As for most

347: sets of singularities the connectedness of the moduli space has

348: not been established (although expected),

349: we state these results in the form of existence.

350: 

351: 

352: \subsection{Contents of the paper}

353: In~\S\ref{S.model}, we use the results of~\cite{symmetric} and

354: construct the \emph{trigonal models} of sextics in question, which

355: are pairs $(\B,\L)$, where $\B$ is a (fixed) trigonal curve

356: in the Hirzebruch surface~$\Sigma_2$ and $\L$ is a (variable)

357: section. We study the conditions on~$\L$ resulting in a particular

358: set of singularities of the sextic. As a consequence, we obtain

359: explicit equations of the sextics and rational parameterizations of

360: the moduli spaces. Theorem~\ref{th.moduli} is proved here.

361: 

362: In~\S\ref{S.vanKampen}, we present the classical Zariski--van

363: Kampen method~\cite{vanKampen} in a form suitable for curves on

364: Hirzebruch surfaces.

365: The contents of this section is a formal account of a few

366: observations found in~\cite{degt.Oka3}

367: and~\cite{degt.kplets}.

368: %It is a more formal and detailed exposition of a

369: %few observations outlined in~\cite{degt.Oka3}

370: %and~\cite{degt.kplets}.

371: 

372: In~\S\ref{S.group}, we apply the classical Zariski--van Kampen

373: theorem to the trigonal models constructed above and obtain

374: presentations of the fundamental groups. The main advantage of

375: this approach (replacing sextics with their trigonal models) is

376: the fact that the number of points to keep track of reduces

377: from~$6$ to~$4$, which simplifies the computation of the braid

378: monodromy. As a first application, we show that all groups can be

379: generated by loops in a small neighborhood of (any) type~$\bE_6$

380: singular point of the curve.

381: 

382: In~\S\ref{S.perturbations}, we study perturbations of

383: sextics considered in~\S\S\ref{S.model} and~\ref{S.group}.

384: We confine ourselves to a few simple cases when the perturbed

385: group is easily found by simple local analysis. This gives $117$

386: new (compared to~\cite{degt.8a2}) sextics with abelian fundamental

387: group

388: %not covered ny Nori's theorem

389: and $8$ sextics of torus type. More complicated

390: perturbations are not necessary, as the resulting sextics are not

391: new, see Remark~\ref{not.new}.

392: 

393: 

394: \section{The trigonal model\label{S.model}}

395: 

396: \subsection{Trigonal curves}

397: Recall

398: that the \emph{Hirzebruch surface}~$\Sigma_2$ is a

399: geometrically ruled rational

400: surface with an exceptional section~$E$ of

401: self-intersection~$(-2)$. A \emph{trigonal curve} is a reduced

402: curve

403: $\B\subset\Sigma_2$ disjoint from~$E$ and intersecting each

404: generic fiber of~$\Sigma_2$ at three points. A \emph{singular

405: fiber} (sometimes referred to as \emph{vertical tangent})

406: of a trigonal curve~$\B$ is a fiber of~$\Sigma_2$ that is

407: not transversal to~$\B$.

408: The double covering~$X$ of~$\Sigma_2$

409: ramified at $\B+E$ is an elliptic surface, and the singular fibers

410: of~$\B$ are the projections of those of~$X$. For this reason, to

411: describe the topological types of singular fibers of~$\B$,

412: we use (one of)

413: the standard notation for the types of singular elliptic fibers,

414: referring to the corresponding extended Dynkin diagrams. The types

415: are as follows:

416: \Dashes

417: \dash

418: $\tilde\bA_0^*$: a simple vertical tangent;

419: \dash

420: $\tilde\bA_0^{**}$: a vertical inflection tangent;

421: \dash

422: $\tilde\bA_1^*$: a node of~$\B$ with one of the branches vertical;

423: \dash

424: $\tilde\bA_2^*$: a cusp of~$\B$ with vertical tangent;

425: \dash

426: $\tilde\bA_p$, $\tilde\bD_q$, $\tilde\bE_6$, $\tilde\bE_7$,

427: $\tilde\bE_8$: a simple singular point of~$\B$ of the same type

428: with minimal possible local intersection index with the fiber.

429: \endDashes

430: For the relation to Kodaira's classification of singular elliptic

431: fibers and further details and references,

432: see~\cite{degt.kplets}. In the

433: present paper, we merely use the notation.

434: 

435: The \emph{\rom(functional\rom) $j$-invariant}

436: $j=j_{\B}\:\Cp1\to\Cp1$ of

437: a trigonal curve $\B\subset\Sigma_2$ is defined as the analytic

438: continuation of the function sending a point

439: $b$ in the base $\Cp1$ of~$\Sigma_2$ representing a

440: nonsingular fiber~$F$ of~$\B$ to the $j$-invariant (divided

441: by~$12^3$)

442: of the elliptic

443: curve covering~$F$ and ramified at $F\cap(\B+E)$. The curve~$\B$

444: is called \emph{isotrivial} if $j_{\B}=\const$. Such curves can

445: easily be enumerated, see, \eg,~\cite{degt.kplets}. The curve~$\B$

446: is called \emph{maximal} if it has the following properties:

447: \Dashes

448: \dash

449: $\B$ has no singular fibers of type~$\bD_4$;

450: \dash

451: $j=j_{\B}$ has no critical values other than~$0$, $1$, and~$\infty$;

452: \dash

453: each point in the pull-back $j^{-1}(0)$ has ramification index at

454: most~$3$;

455: \dash

456: each point in the pull-back $j^{-1}(1)$ has ramification index at

457: most~$2$.

458: \endDashes

459: The maximality of a non-isotrivial trigonal curve $\B\subset\Sigma_2$

460: can easily be detected by

461: applying the Riemann--Hurwitz formula to the map

462: $j_{\B}\:\Cp1\to\Cp1$; it depends only

463: on the (combinatorial) set of singular fibers of~$\B$,

464: see~\cite{degt.kplets} for details. The classification of such

465: curves reduces to a combinatorial problem;

466: a partial classification of

467: maximal trigonal curves in~$\Sigma_2$ is found

468: in~\cite{symmetric}. An important property of maximal trigonal

469: curves is their rigidity, see~\cite{degt.kplets}: any small

470: deformation of such a curve~$\B$ is isomorphic to~$\B$. For this

471: reason, we do not need to keep parameters in the equations below.

472: 

473: \subsection{The trigonal curve~$\B$}\label{s.trigonal}

474: Let~$B$ be an \emph{irreducible} sextic of torus type with

475: \emph{simple singularities only} and with at least two

476: type~$\bE_6$ singular point. (Below, we show that the emphasized

477: properties hold automatically, see~\ref{proof.moduli}.)

478: Clearly,

479: the set of inner singularities of~$B$ can only be $(3\bE_6)$,

480: $(2\bE_6\splus\bA_5)$, or $(2\bE_6\splus2\bA_2)$. Hence,

481: according to~\cite{symmetric}, $B$

482: has an involutive symmetry (\ie, projective automorphism)~$c$

483: stable under equisingular deformations. Let~$L_c$ and~$O_c$ be,

484: respectively, the fixed line and the isolated fixed point of~$c$.

485: One has $O_c\notin B$.

486: Denote

487: by $\Cp2(O_c)$ the blow-up of~$\Cp2$ at~$O_c$. Then,

488: the quotient $\Cp2(O_c)/c$ is the

489: Hirzebruch surface~$\Sigma_2$

490: %, \ie, geometrically ruled rational

491: %surface with an exceptional section~$E$ of self-intersection~$-2$,

492: and the projection $B/c$ is a trigonal curve $\B\subset\Sigma_2$

493: %(disjoint from~$E$)

494: with the set of singularities

495: $\bE_6\splus\bA_2$. The double covering $\Cp2(O_c)\to\Sigma_2$ is

496: ramified at~$E$ and a generic section $\L\subset\Sigma_2$ (the image

497: $L_c/c$) disjoint from~$E$ and not passing through the

498: type~$\bE_6$ singular point of~$\B$ (as otherwise the two

499: type~$\bE_6$ singular points of~$B$ would merge to a single

500: non-simple singularity).

501: 

502: Conversely, given a trigonal curve $\B\subset\Sigma_2$ with the

503: set of singularities $\bE_6\splus\bA_2$ and a section

504: $\L\subset\Sigma_2$ disjoint from~$E$ and not passing through the

505: type~$\bE_6$ singular point of~$\B$, the pull-back of~$\B$ in the

506: double covering of $\Sigma_2/E$ ramified at $E/E$ and~$\L$ is a

507: sextic $B\subset\Cp2$ with at least two type~$\bE_6$ singular

508: points. Below we show that $B$ is necessarily of torus type,

509: see~\eqref{eq.torus}.

510: 

511: 

512: \subsection{Equations}\label{s.equations}

513: Any trigonal curve $\B\subset\Sigma_2$ with the set of

514: singularities $\bE_6\splus\bA_2$ is either isotrivial or maximal

515: (see~\cite{symmetric} for precise definitions); in particular,

516: such curves are rigid, \ie, within each of the two families, any

517: two curves are isomorphic in~$\Sigma_2$. A curve~$\B$ can be

518: obtained by an elementary transformation from a cuspidal cubic

519: $C\subset\Sigma_1=\Cp2(O)$: the blow-up center~$O$ should be chosen

520: on the inflection tangent to~$C$, and the elementary

521: transformation should contract this tangent.

522: 

523: In appropriate affine coordinates $(x,y)$ in~$\Sigma_2$ any

524: trigonal

525: curve~$\B$ as above can be given by an equation of the form

526: $$

527: f_r(x,y):=y^3+r^2y^2+2rxy+x^2=0,

528: \eqtag\label{eq.equation}

529: $$

530: where $r\in\C$ is a parameter. If $r=0$, the curve is isotrivial,

531: its $j$-invariant being $j\equiv0$. Otherwise, the automorphism

532: $(x,y)\mapsto(r^3x,r^2y)$ of~$\Sigma_2$ converts the curve to

533: $f_1(x,y)=0$. Below, in all plots and numeric evaluation, we use

534: the value $r=3$.

535: 

536: The

537: %discriminant with respect to~$y$

538: $y$-discriminant

539: of the polynomial~$f_r$ given

540: by~\eqref{eq.equation} is $-x^3(27x-4r^3)$. Thus, if $r\ne0$, the

541: curve has three singular fibers, of types~$\tilde\bA_2$,

542: $\tilde\bA_0^{*}$ (vertical tangent),

543: and~$\tilde\bE_6$ over $x=0$, $4r^3\!/27$,

544: and~$\infty$, respectively. In the isotrivial case $r=0$, there

545: are two singular fibers, of types~$\tilde\bA_2^*$

546: and~$\tilde\bE_6$, over $x=0$ and~$\infty$, respectively.

547: 

548: The curve~$\B$ is rational; it can be parameterized by

549: $$

550: x=\xt:=rt^2+t^3,\qquad y=\yt:=-t^2.

551: \eqtag\label{eq.parameterization}

552: $$

553: The vertical tangency point of~$\B$ corresponds to the value

554: $t=-2r/3$.

555: 

556: Consider a section~$\L$ of~$\Sigma_2$ given by

557: $$

558: y=s(x):=ax^2+bx+c,\quad a\ne0.

559: \eqtag\label{eq.section}

560: $$

561: (The assumption $a\ne0$ is due to the fact that $\L$ should not

562: pass through the type~$\bE_6$ singular point of~$\B$.) Let

563: $B\subset\Cp2$ be the pull-back of~$\B$ under the double covering

564: of $\Sigma_2/E$ ramified at $E/E$ and~$\L$. It is a plane sextic

565: which, in appropriate affine coordinates $(x,y)$ in~$\Cp2$, is

566: given by the equation

567: $$

568: f_r(x, y^2+s(x))=0.

569: \eqtag\label{eq.sextic}

570: $$

571: Obviously, $B$ is of torus type, the torus structure being

572: $$

573: f_r(x,\bar y)=\bar y^3+(r\bar y+x)^2,\quad \bar y=y^2+s(x).

574: %(y^2-s(x))^3+(ry^2-rs(x)+x)^2=0.

575: \eqtag\label{eq.torus}

576: $$

577: According to~\cite{degt.Oka}, this is the only torus structure

578: on~$B$. The inner singularities of~$B$ are two type~$\bE_6$ points

579: over the type~$\bE_6$ point of~$\B$ and two cusps or one

580: type~$\bA_5$ or~$\bE_6$ point over the cusp of~$\B$. (There is

581: only one point if $\L$ passes through the cusp of~$\B$; this point

582: is of type~$\bE_6$ if $\L$ is tangent to~$\B$ at the cusp.) The

583: outer singularities of~$B$ arise from the tangency of~$\L$

584: and~$\B$: each point of $p$-fold intersection, $p>1$, of~$\L$

585: and~$\B$ smooth for~$\B$ gives rise to a type~$\bA_{p-1}$ outer

586: singularity of~$B$. For detail, see~\cite{degt.Oka3}.

587: 

588: In the rest of this section, we discuss various degenerations of

589: the pair $(\B,\L)$ and parameterize the corresponding triples

590: $(a,b,c)$. For convenience, each time we mention parenthetically

591: the set of singularities of

592: the sextic~$B$ arising from $(\B,\L)$.

593: 

594: \subsection{Tangents and double tangents}\label{s.tangent}

595: Equating the values and the derivatives of $s(\xt(t))$

596: and~$\yt(t)$, one concludes that

597: a section~$\L$ as in~\eqref{eq.section}

598: is tangent to~$\B$ at a point $(\xt(t),\yt(t))$,

599: $t\ne0$, $-2r/3$, (the set of singularities

600: $(2\bE_6\splus2\bA_2)\splus\bA_1$)

601: if and only if

602: %$$

603: %\gathered

604: %b=-\frac{2(3at^4+5art^3+2ar^2t^2+1)}{3t+2r},\\

605: %c=\frac{t^3(3at^4+8art^3+7ar^2t^2+2ar^3t-1)}{3t+2r}.

606: %\endgathered

607: %$$

608: $$

609: b=-2t^2(t+r)a-\frac2{3t+2r},\quad

610: c=t^4(t+r)^2a-\frac{t^3}{3t+2r}.

611: \eqtag\label{eq.tangent}

612: $$

613: 

614: Double tangents are described by the following lemma.

615: 

616: \lemma\label{double.tangent}

617: There exists a section~$\L$ tangent to the curve~$\B$

618: at two distinct points

619: $(\xt(t_1),\yt(t_1))$ and $(\xt(t_2),\yt(t_2))$, $t_1\ne t_2$, if

620: and only if $t_1+t_2=-r/3$ and neither~$t_1$ nor~$t_2$ is~$0$,

621: $-r/6$, or $-2r/3$.

622: \endlemma

623: 

624: \proof

625: Substituting $t=t_1$ and $t=t_2$ to~\eqref{eq.tangent}, equating

626: the resulting values of~$b$ and~$c$, solving both equations

627: for~$a$, and equating the results, one obtains

628: $(t_1-t_2)^2(3t_1+3t_2+r)=0$; now, the statement is immediate.

629: \endproof

630: 

631: Thus, a section~$\L$ as in~\eqref{eq.section} is double tangent

632: to~$\B$ (the set of singularities

633: $(2\bE_6\splus2\bA_2)\splus2\bA_1$)

634: if and only if, for some $t\ne0$, $-r/6$, $-2r/3$, one has

635: $$

636: \gathered

637: a=-\frac{27}{(3t-r)^2(3t+2r)^2},\\

638: b=\frac{2r(27t^2+9rt-2r^2)}{(3t-r)^2(3t+2r)^2},\\

639: c=-\frac{2t^3(3t+r)^3}{(3t-r)^2(3t+2r)^2}.

640: \endgathered

641: \eqtag\label{eq.double}

642: $$

643: 

644: A point of quadruple intersection of~$\L$ and~$\B$ can be obtained

645: from Lemma~\ref{double.tangent} letting $t_1=t_2$. (Alternatively,

646: one can equate the derivatives of order~$0$ to~$3$ of $s(\xt(t))$

647: and $\yt(t)$.) As a result, $(\xt(t),\yt(t))$ is a point of

648: quadruple intersection of~$\L$ and~$\B$ (the set of singularities

649: $(2\bE_6\splus2\bA_2)\splus\bA_3$) if and only if

650: $$

651: t=-\frac{r}6,\qquad

652: (a,b,c)=\Bigl(-\frac{16}{3r^4},-\frac{88}{81r},\frac{r^2}{4374}\Bigr).

653: \eqtag\label{eq.2e6+2a2+a3}

654: $$

655: All points of intersection of this section~$\L$ and~$\B$ are:

656: \Dashes

657: \dash

658: transversal intersection at

659: $t=\Bigl(-\dfrac23+\dfrac{\sqrt2}2\Bigr)r$,

660: $x=\Bigl(-\dfrac{19}{54}+\dfrac{\sqrt2}4\Bigr)r^3

661: \approx.0459$;

662: \dash

663: transversal intersection at

664: $t=\Bigl(-\dfrac23-\dfrac{\sqrt2}2\Bigr)r$,

665: $x=\Bigl(-\dfrac{19}{54}-\dfrac{\sqrt2}4\Bigr)r^3

666: \approx-19.1$;

667: \dash

668: quadruple intersection at $t=-\dfrac{r}6$,

669: $x=\dfrac{5r^3}{216}=.625$.

670: \endDashes

671: The curve~$\B$ and the section~$\L$ given by~\eqref{eq.2e6+2a2+a3}

672: are plotted in Figure~\ref{fig.2e6+2a2+a3} (in black and grey,

673: respectively). The section is above the curve over $x=0$; it

674: intersects the topmost branch over $x\approx.0459$ and is tangent

675: to the middle branch over $x=.625$.

676: 

677: \midinsert

678: \plot{2e6+2a2+a3}

679: \figure

680: The set of singularities $(2\bE_6\splus2\bA_2)\splus\bA_3$

681: \endfigure\label{fig.2e6+2a2+a3}

682: \endinsert

683: 

684: 

685: \subsection{Sections through the cusp}\label{s.cusp}

686: A section~$\L$ as in~\eqref{eq.section}

687: passes through the cusp of~$\B$ (the set of

688: singularities $(2\bE_6\splus\bA_5)$)

689: if and only if $c=0$;

690: it is tangent to~$\B$ at the cusp (the set of singularities

691: $(3\bE_6)$)

692: if and only if, in addition,

693: $b=-1/r$.

694: 

695: A section tangent to~$\B$ at a point $(\xt(t),\yt(t))$,

696: see~\eqref{eq.tangent},

697: passes through the cusp of~$\B$ (the set of

698: singularities $(2\bE_6\splus\bA_5)\splus\bA_1$)

699: if and only if

700: $$

701: a=\frac1{t(t+r)^2(3t+2r)},\quad

702: b=-\frac{2(2t+r)}{(t+r)(3t+2r)},\quad

703: c=0,

704: \eqtag\label{eq.cusp.tangent}

705: $$

706: $t\ne0$, $-r$, $-2r/3$. (Note that the value $t=-r$ corresponds to

707: the smooth point of~$\B$ in the same vertical fiber as the cusp.)

708: Such a section is tangent to~$\B$ at the cusp (the set of

709: singularities $(3\bE_6)\splus\bA_1$) if and only if

710: $$

711: t=-\frac{r}3,\qquad

712: (a,b,c)=\Bigl(-\frac{27}{4r^4},-\frac1r,0\Bigr).

713: \eqtag\label{eq.3e6+a1}

714: $$

715: The points of intersection of the latter section~$\L$ and~$\B$

716: are:

717: \Dashes

718: \dash

719: the cusp of~$\B$ at $t=0$, $x=0$;

720: \dash

721: transversal intersection at $t=-\dfrac{4r}3$,

722: $x=-\dfrac{16r^3}{27}=-16$;

723: \dash

724: tangency at $t=-\dfrac{r}3$, $x=\dfrac{2r^3}{27}=2$.

725: \endDashes

726: The section~$\L$ given by~\eqref{eq.3e6+a1} looks similar to that

727: shown in Figure~\ref{fig.2e6+2a2+a3}. (Near the cusp of~$\B$, the

728: two curves are too close to be distinguished visually.) Between

729: $x=0$ and $x=2$, the section lies between the topmost and middle

730: branches of~$\B$.

731: 

732: 

733: \subsection{Inflection tangents}\label{s.inflection}

734: Equating the derivatives of $s(\xt(t))$ and $\yt(t)$ of order $0$,

735: $1$, and~$2$, one can see that

736: a section~$\L$ as in~\eqref{eq.section}

737: is inflection tangent to~$\B$ at a

738: point $(\xt(t),\yt(t))$, $t\ne0$, $-2r/3$, (the set of

739: singularities $(2\bE_6\splus2\bA_2)\splus\bA_2$) if and only if

740: $$

741: a=\frac3{t(3t+2r)^3},\quad

742: b=-\frac{2(12t^2+15rt+4r^2)}{(3t+2r)^3},\quad

743: c=-\frac{t^3(6t^2+6rt+r^2)}{(3t+2r)^3}.

744: \eqtag\label{eq.inflection}

745: $$

746: Such a section passes through the cusp of~$\B$ (the set of

747: singularities $(2\bE_6\splus\bA_5)\splus\bA_2$) if and only if

748: $t=(-3\pm\sqrt3)r/6$. Thus, we obtain \emph{two families}, which

749: are Galois conjugate over $\Q[\sqrt3]$, \cf.~\cite{OkaPho.moduli}.

750: For one of the families, one has

751: $$

752: t=\Bigl(-\frac12+\frac{\sqrt3}6\Bigr)r,\quad

753: (a,b,c)=\Bigl(\frac{12(3-2\sqrt3)}{r^4},-\frac{4(2-\sqrt3)}r,0\Bigr),

754: \eqtag\label{eq.2e6+a5+a2.1}

755: $$

756: and the points of intersection of~$\L$ and~$\B$ are:

757: \Dashes

758: \dash

759: the cusp of~$\B$ at $t=0$, $x=0$;

760: \dash

761: transversal intersection at

762: $t=\Bigl(-\dfrac12-\dfrac{\sqrt3}2\Bigr)r$,

763: $x=\Bigl(-\dfrac14-\dfrac{\sqrt3}4\Bigr)r^3\approx-18.4$;

764: \dash

765: inflection tangency at $t=\Bigl(-\dfrac12+\dfrac{\sqrt3}6\Bigr)r$,

766: $x=\Bigl(\dfrac1{12}-\dfrac{\sqrt3}{36}\Bigr)r^3\approx.951$.

767: \endDashes

768: This section looks similar to that shown in

769: Figure~\ref{fig.2e6+2a2+a3}; between $x=0$ and $x\approx.951$, the

770: section is just below the middle branch of the curve.

771: 

772: For the other family, one has

773: $$

774: t=\Bigl(-\frac12-\frac{\sqrt3}6\Bigr)r,\quad

775: (a,b,c)=\Bigl(\frac{12(3+2\sqrt3)}{r^4},-\frac{4(2+\sqrt3)}r,0\Bigr),

776: \eqtag\label{eq.2e6+a5+a2.2}

777: $$

778: and the points of intersection of~$\L$ and~$\B$ are:

779: \Dashes

780: \dash

781: the cusp of~$\B$ at $t=0$, $x=0$;

782: \dash

783: transversal intersection at

784: $t=\Bigl(-\dfrac12+\dfrac{\sqrt3}2\Bigr)r$,

785: $x=\Bigl(-\dfrac14+\dfrac{\sqrt3}4\Bigr)r^3\approx4.94$;

786: \dash

787: inflection tangency at $t=\Bigl(-\dfrac12-\dfrac{\sqrt3}6\Bigr)r$,

788: $x=\Bigl(\dfrac1{12}+\dfrac{\sqrt3}{36}\Bigr)r^3\approx3.55$.

789: \endDashes

790: The curve~$\B$ and the section~$\L$ given by~\eqref{eq.2e6+a5+a2.2}

791: are plotted in Figure~\ref{fig.2e6+a5+a2}, in black and grey,

792: respectively.

793: 

794: \midinsert

795: \plot{2e6+a5+a2}

796: \figure

797: The set of singularities

798: $(2\bE_6\splus\bA_5)\splus\bA_2$, the family~\eqref{eq.2e6+a5+a2.2}

799: \endfigure\label{fig.2e6+a5+a2}

800: \endinsert

801: 

802: \subsection{Proof of Theorem~\ref{th.moduli}}\label{proof.moduli}

803: First, note that any

804: sextic with two type~$\bE_6$ singular points

805: is irreducible and has simple singularities only. The first

806: statement follows from the fact that an irreducible curve of

807: degree~$4$ or~$5$ (respectively,~$\le3$) may have at most one

808: (respectively, none) type~$\bE_6$ singular point, and the second

809: one, from the fact that a type~$\bE_6$ (respectively, non-simple)

810: singular point takes~$3$ (respectively,~$\ge6$) off the

811: genus, whereas the genus of a nonsingular sextic is~$10$. Thus, we

812: can apply the results of~\cite{symmetric} enumerating stable

813: symmetries of curves.

814: 

815: For a set of

816: singularities~$\Sigma\supset2\bE_6$, consider the moduli

817: space $\CM(\Sigma)$ of sextics~$B$ of torus type with the set of

818: singularities~$\Sigma$ and

819: the moduli space $\tCM(\Sigma)$ of pairs $(B,c)$, where $B$

820: is a sextic as above and $c$ is a

821: stable involution of~$B$. Due to~\cite{symmetric}, the forgetful

822: map $\tCM(\Sigma)\to\CM(\Sigma)$ is generically finite-to-one and

823: onto.

824: 

825: As explained in Sections~\ref{s.trigonal} and~\ref{s.equations},

826: the space $\tCM(\Sigma)$ can be identified with the moduli space

827: of pairs $(\B,\L)$, where $\B\subset\Sigma_2$ is a trigonal curve

828: given by~\eqref{eq.equation} and $\L$ is a section of~$\Sigma_2$

829: in a certain prescribed position with respect to~$\B$. The spaces

830: of pairs $(\B,\L)$ are described in

831: Sections~\ref{s.tangent}--\ref{s.inflection}, and for each

832: $\Sigma\ne(2\bE_6\splus\bA_5)\splus\bA_2$, an explicit rational

833: parameterization is found. (Strictly speaking, in order to pass to

834: the moduli, we need to fix a value of~$r$, say, $r=3$. This

835: results in a Zariski open subset of the moduli space. The portion

836: corresponding to $r=0$ has positive codimension as the isotrivial

837: curve $f_0=0$ has $1$-dimensional group~$\C^*$ of symmetries.)

838: Hence, the space $\tCM(\Sigma)$ is rational and, if

839: $\dim\CM(\Sigma)\le2$, so is $\CM(\Sigma)$. The only case when

840: $\dim\CM(\Sigma)\ge3$ is $\Sigma=(2\bE_6\splus2\bA_2)$. In this

841: case, each curve~$B$ has a unique stable involution,

842: see~\cite{symmetric}, and the map $\tCM(\Sigma)\to\CM(\Sigma)$ is

843: generically one-to-one; hence, $\CM(\Sigma)$ is still rational.

844: 

845: In the exceptional case $\Sigma=(2\bE_6\splus\bA_5)\splus\bA_2$, the

846: space $\CM(\Sigma)=\tCM(\Sigma)$ consists of two points. The fact

847: that any sextic with this set of singularities is of torus type

848: follows immediately from~\cite{JAG}.

849: \qed

850: 

851: \Remark

852: The only sets of singularities containing $2\bE_6$

853: where the curves have more than one

854: (three) stable involutions are $(3\bE_6)$ and $(3\bE_6)\splus\bA_1$,

855: see~\cite{symmetric}. In both cases, the group of stable

856: symmetries can be identified with the group~$\SG3$ of permutations

857: of the three type~$\bE_6$ points. It follows that all three

858: involutions are conjugate by stable symmetries; hence, the map

859: $\tCM(\Sigma)\to\CM(\Sigma)$ is still one-to-one.

860: \endRemark

861: 

862: 

863: \section{Van Kampen's method in Hirzebruch surfaces}\label{S.vanKampen}

864: 

865: In this section, we give a formal and

866: detailed exposition of a few

867: observations outlined in~\cite{degt.Oka3}. Keeping in mind future

868: applications, we treat the general case of a Hirzebruch

869: surface~$\Sigma_k$, $k\ge1$, and a \emph{$d$-gonal curve}

870: $C\subset\Sigma_k$, see Definition~\ref{def.d-gonal}.

871: 

872: Certainly, the essence of this approach is due to van

873: Kampen~\cite{vanKampen}; we merely introduce a few restrictions to

874: the objects used in the construction which make the choices

875: involved slightly more canonical and easier to handle. By no means

876: do we assert that the restrictions are necessary for the approach

877: to work in general.

878: 

879: \subsection{Preliminary definitions}\label{s.defs}

880: Fix a Hirzebruch surface~$\Sigma_k$, $k\ge1$. Denote by

881: $p\:\Sigma_k\to\Cp1$ the ruling, and let $E\subset\Sigma_k$ be

882: the exceptional section, $E^2=-k$. Given a point~$b$ in the

883: base~$\Cp1$, we denote by~$F_b$ the fiber $p^{-1}(b)$. Let

884: $F_b^\circ$ be the `open fiber' $F_b\sminus E$. Observe that

885: $F_b^\circ$ is a dimension~$1$ affine space over~$\C$; hence, one

886: can speak about lines, circles, convexity, convex hulls, \etc.

887: in~$F_b^\circ$.

888: (Thus, strictly speaking, the notation~$F_b^\circ$ means slightly

889: more than just the set theoretical difference $F_b\sminus E$: we

890: always consider~$F_b^\circ$ with its canonical affine structure.)

891: Define the \emph{convex hull} $\Conv C$

892: of a subset

893: $C\subset\Sigma_k\sminus E$ as the union of its fiberwise convex

894: hulls:

895: $$

896: \Conv C=\bigcup_{b\in\Cp1}\Conv(C\cap F_b^\circ).

897: $$

898: 

899: \definition\label{def.d-gonal}

900: Let $d\ge1$ be an integer.

901: A \emph{$d$-gonal curve} (or \emph{degree~$d$ curve})

902: on~$\Sigma_k$ is a reduced algebraic

903: curve $C\in\ls|dE+dkF|$ disjoint from

904: the exceptional section~$E$. (Here, $F$ is any fiber

905: of~$\Sigma_k$.) A \emph{singular fiber} of a $d$-gonal curve~$C$

906: is a fiber of~$\Sigma_k$ that intersects~$C$ at fewer than $d$

907: points. (With a certain abuse of the language, the points in the

908: base $\Cp1$ whose pull-backs are singular fibers will also be

909: referred to as singular fibers of~$C$.)

910: \enddefinition

911: 

912: \Remark\label{affine}

913: Recall that the complement $\Sigma_x\sminus E$ can be covered by

914: two affine charts, with coordinates $(x,y)$ and $(x',y')$ and

915: transition function $x'=1/x$, $y'=y/x^k$. In the coordinates

916: $(x,y)$, any $d$-gonal curve~$C$ is given by an equation of the form

917: $$

918: f(x,y)=\sum_{i=0}^dy^iq_{i}(x)=0,\quad

919:  \deg q_i=k(d-i),\quad

920:  q_d=\const\ne0,

921: $$

922: and the singular fibers of~$C$ are those of the form $F_x$, where $x$ is

923: a root of the $y$-discriminant~$D_y$ of~$f$. (The fiber $F_\infty$ over

924: $x=\infty$ is singular for~$C$ if and only if $\deg D_y<kd(d-1)$.)

925: \endRemark

926: 

927: 

928: \subsection{Proper sections and braid monodromy}\label{s.proper}

929: Fix a $d$-gonal curve $C\subset\Sigma_k$.

930: 

931: \definition\label{def.proper}

932: Let $\disk\subset\Cp1$ be a closed (topological) disk. A partial

933: section $s\:\disk\to\Sigma_k$ of~$p$ is called \emph{proper} if its

934: image is disjoint from both~$E$ and $\Conv C$.

935: \enddefinition

936: 

937: \lemma\label{proper.homotopic}

938: Any disk $\disk\subset\Cp1$ admits a proper section

939: $s\:\disk\to\Sigma_k$.

940: Any two proper sections over~$\disk$ are homotopic

941: in the class of proper sections\rom;

942: furthermore, any homotopy over a fixed

943: point $b\in\disk$ extends to a homotopy over~$\disk$.

944: \endlemma

945: 

946: \proof

947: The restriction~$p'$ of~$p$ to $\Sigma_k\sminus(E\cup\Conv C)$ is a

948: locally trivial fibration with a typical fiber~$F'$

949: homeomorphic to a punctured open disk. Since $\disk$ is

950: contractible, $p'$ is trivial over~$\disk$ and, after trivializing,

951: sections over~$\disk$ can be identified with maps $\disk\to F'$.

952: Such maps do exist, and

953: any two such maps are homotopic, again due to the fact that

954: $\disk$ is contractible.

955: \endproof

956: 

957: Pick a closed disk $\disk\subset\Cp1$ as above and denote

958: $\disk^\sharp=\disk\sminus\{b_1,\ldots,b_l\}$,

959: where $b_1,\ldots,b_l$ are the singular fibers of~$C$ that belong

960: to~$\disk$.

961: Fix a point

962: $b\in\disk^\sharp$. The restriction

963: $p^\sharp\:p^{-1}(\disk^\sharp)\sminus(C\cup E)\to\disk^\sharp$

964: is a locally trivial fibration with a typical fiber

965: $F_b^\circ\sminus C$, and any proper section $s\:\disk\to\Sigma_k$

966: restricts to

967: a section of~$p^\sharp$. Hence, given a proper

968: section~$s$, one can define the group

969: $\pi_F:=\pi_1(F_b^\circ\sminus C,s(b))$ and the \emph{braid

970: monodromy} $m\:\pi_1(\disk^\sharp,b)\to\Aut\pi_F$. Informally, for

971: a loop $\Gs\:[0,1]\to\disk^\sharp$, the automorphism $m([\Gs])$

972: of~$\pi_F$ is

973: obtained by dragging the fiber~$F_b$ along $\Gs(t)$ while keeping the

974: base point on $s(\Gs(t))$. (Formally, it is obtained by

975: trivializing the fibration $\Gs^*p^\sharp$.)

976: 

977: It is essential that, in this paper,

978: %{\it

979: we reserve the term

980: `braid monodromy' for the homomorphism~$m$ constructed using a

981: \emph{proper} section~$s$.

982: %}

983: Under this convention,

984: the following

985: lemma is an immediate consequence of

986: Lemma~\ref{proper.homotopic} and the obvious fact that the braid

987: monodromy is homotopy invariant.

988: 

989: \lemma\label{proper.braid}

990: The braid monodromy $m\:\pi_1(\disk^\sharp,b)\to\Aut\pi_F$ is well

991: defined and independent of the choice of a proper section

992: over~$\disk$ passing through $s(b)$.

993: \qed

994: \endlemma

995: 

996: \Remark

997: More generally, given a path

998: $\tilde\Gs\:[0,1]\to p^{-1}(\disk^\sharp)\sminus(\Conv C\cup E)$,

999: %and letting $\Gs=p\circ\tilde\Gs\:[0,1]\to\disk^\sharp$,

1000: one can use Lemma~\ref{proper.homotopic} to conclude that the

1001: braid monodromy commutes with the translation isomorphism

1002: $$

1003: T_{\Gs}\:\pi_1(\disk^\sharp,\Gs(0))\to\pi_1(\disk^\sharp,\Gs(1))

1004: $$

1005: (where $\Gs=p\circ\tilde\Gs\:[0,1]\to\disk^\sharp$)

1006: and the isomorphism

1007: $$

1008: \Aut\pi_1(F_{\Gs(0)}^\circ\sminus C,\tilde\Gs(0))\to

1009:  \Aut\pi_1(F_{\Gs(1)}^\circ\sminus C,\tilde\Gs(1))

1010: $$

1011: induced by the translation $T_{\tilde\Gs}$ along~$\tilde\Gs$.

1012: \endRemark

1013: 

1014: \Remark\label{constant.section}

1015: For most computations, we will take for~$s$ a `constant section'

1016: constructed as follows: pick an affine coordinate system $(x,y)$,

1017: see Remark~\ref{affine}, so that the point $x=\infty$ does

1018: \emph{not} belong to~$\disk$, and let~$s$ be the section

1019: $x\mapsto c=\const$, $\ls|c|\gg0$. (In other words, the graph

1020: of~$s$ is the $1$-gonal curve $\{y=c\}\subset\Sigma_k$.) Since the

1021: intersection $p^{-1}(\disk)\cap\Conv C\subset\Sigma_k\sminus E$ is

1022: compact, such a section is indeed proper whenever $\ls|c|$ is

1023: sufficiently large.

1024: \endRemark

1025: 

1026: \Remark\label{extension}

1027: Another consequence of Lemma~\ref{proper.braid} is the fact

1028: that, for

1029: any nested pair of disks $\disk_1\subset\disk_2$, the braid

1030: monodromy commutes with the inclusion homomorphism

1031: $\pi_1(\disk_1^\sharp)\to\pi_1(\disk_2^\sharp)$. Indeed, one can

1032: construct both

1033: %braid

1034: monodromies using a proper section

1035: over~$\disk_2$ and restricting it to~$\disk_1$ when necessary.

1036: \endRemark

1037: 

1038: Pick a basis $\zeta_1,\ldots,\zeta_d$ for $\pi_F$ and a basis

1039: $\Gs_1,\ldots,\Gs_l$ for $\pi_1(\disk^\sharp,b)$. Denote

1040: $m_i=m(\Gs_i)$, $i=1,\ldots,l$.

1041: The following

1042: statement is the essence of Zariski--van Kampen's method for

1043: computing the fundamental group of a plane algebraic curve,

1044: see~\cite{vanKampen} for the proof and further details.

1045: 

1046: \theorem\label{th.vanKampen.original}

1047: Let $\disk\subset\Cp1$ be a closed disk as above,

1048: and assume that the boundary

1049: $\partial\disk$ is free of singular fibers of~$C$. Then one has

1050: $$

1051: \pi_1(p^{-1}(\disk)\sminus(C\cup E),s(b))=

1052:  \bigl<\zeta_1,\ldots,\zeta_d\bigm|m_i=\id,\ i=1,\ldots,l\bigl>,

1053: $$

1054: where each \emph{braid relation} $m_i=\id$ should be understood as

1055: a $d$-tuple of relations $\zeta_j=m_i(\zeta_j)$, $j=1,\ldots,d$.

1056: \qed

1057: \endtheorem

1058: 

1059: 

1060: \subsection{The monodromy at infinity}\label{s.monodromy.infty}

1061: Let $b\in\disk^\sharp\subset\disk\subset\Cp1$ be as in

1062: Section~\ref{s.proper}. Denote by $\Gr_b\in\pi_F$ the

1063: `counterclockwise' generator of the abelian subgroup

1064: $\Z\cong\pi_1(F_b^\circ\sminus\Conv C)\subset\pi_F$. (In other

1065: words, $\Gr_b$ is the class of a large circle in~$F_b^\circ$

1066: encompassing $\Conv C\cap F_b^\circ$.

1067: If

1068: $\zeta_1,\ldots,\zeta_d$ is a `standard basis' for~$\pi_F$, \cf.

1069: Figure~\ref{fig.basis}, left, then

1070: $\Gr_b=\zeta_1\cdot\ldots\cdot\zeta_d$.) Clearly, $\Gr_b$ is invariant

1071: under the braid monodromy and, properly understood, it is

1072: preserved by the translation homomorphism along any path in

1073: $p^{-1}(\disk^\sharp)\sminus(\Conv C\cup E)$. (Indeed,

1074: as explained in the proof of Lemma~\ref{proper.homotopic},

1075: the fibration

1076: $p^{-1}(\disk)\sminus(\Conv C\cup E)\to\disk$ is trivial,

1077: hence $1$-simple.) Thus, there is a canonical identification of

1078: the elements $\Gr_{b'}$, $\Gr_{b''}$ in the fibers over any two

1079: points $b',b''\in\disk^\sharp$; for this reason, we will omit the

1080: subscript~$b$ in the sequel.

1081: 

1082: Assume that the boundary $\partial\disk$ is free of singular

1083: fibers of~$C$. Then, connecting $\partial\disk$ with the base

1084: point~$b$ by a path in~$\disk^\sharp$ and traversing it in the

1085: counterclockwise direction (with respect to the canonical complex

1086: orientation of~$\disk$), one obtains a certain element

1087: $[\partial\disk]\in\pi_1(\disk^\sharp,b)$ (which depends

1088: on the choice of the path above).

1089: 

1090: \proposition\label{monodromy.infty}

1091: In the notation above, assume that the interior of~$\disk$

1092: contains all singular fibers of~$C$. Then, for any

1093: $\zeta\in\pi_F$, one has

1094: $m([\partial\disk])(\zeta)=\Gr^{k}\zeta\Gr^{-k}$. \rom(In

1095: particular, $m([\partial\disk])$ does not depend on the choices in

1096: the definition of $[\partial\disk]$.\rom)

1097: \endproposition

1098: 

1099: \proof

1100: Due to the homotopy invariance of the braid monodromy (and the

1101: invariance of~$\Gr$), one can replace~$\disk$ with any larger disk

1102: and assume that the base point~$b$ is in the boundary.

1103: Consider affine charts $(x,y)$ and $(x',y')$, see

1104: Remark~\ref{affine},

1105: such that the fiber

1106: $\{x=\infty\}=\{x'=0\}$ does not belong to~$\disk$ (and hence is

1107: nonsingular for~$C$), and replace $\disk$ with the disk

1108: $\{\ls|x|\le1/\epsilon\}$ for some positive $\epsilon\ll1$.

1109: About $x'=0$, the curve~$C$ has $d$ analytic branches of the form

1110: $y'=c_i+x'\Gf_i(x')$,

1111: where $c_i$ are pairwise distinct constants and $\Gf_i$ are

1112: analytic functions, $i=1,\ldots,d$. Restricting these expressions

1113: to the circle $x'=\epsilon\exp(-2\pi t)$, $t\in[0,1]$,

1114: and passing to

1115: $x=1/x'$ and $y=y'x^k$, one obtains

1116: $y=c_i\epsilon^{-k}\exp(2k\pi t)+O(\epsilon^{-k+1})$,

1117: $i=1,\ldots,d$. Thus,

1118: from the point of view of a trivialization of the ruling

1119: over~$\disk$ (\eg, the one given by~$y$),

1120: the parameter~$\epsilon$ can be chosen so small that

1121: %if $\epsilon$ is small enough,

1122: the $d$

1123: branches move along $d$ pairwise disjoint concentric circles

1124: (not quite round), each

1125: branch making $k$ turns in the counterclockwise direction. On the

1126: other hand, one can assume that

1127: the base point remains in a constant section $y=c=\const$

1128: with $\ls|c|\gg\epsilon^{-k}\max\ls|c_i|$, see

1129: Remark~\ref{constant.section}. The resulting braid is the

1130: conjugation by~$\Gr^{-k}$.

1131: \endproof

1132: 

1133: \subsection{The relation at infinity}

1134: We are ready to state the principal result of this section.

1135: Fix a $d$-gonal curve $C\subset\Sigma_k$ and

1136: choose a closed disk $\disk\subset\Cp1$ satisfying the following

1137: conditions:

1138: \roster

1139: \item\local{all.but.1}

1140: $\disk$ contains all but at most one singular fibers of~$C$;

1141: \item\local{no.boundary}

1142: none of the singular fibers of~$C$ is in the boundary

1143: $\partial\disk$.

1144: \endroster

1145: As in Section~\ref{s.proper}, pick a base point

1146: $b\in\disk^\sharp$, a basis $\zeta_1,\ldots,\zeta_d$ for the group

1147: $\pi_F$ over~$b$, and a basis $\Gs_1,\ldots,\Gs_l$ for the group

1148: $\pi_1(\disk^\sharp,b)$. Let $m_i=m(\Gs_i)$, $i=1,\ldots,l$, where

1149: $m\:\pi_1(\disk^\sharp,b)\to\Aut\pi_F$ is the braid monodromy.

1150: 

1151: \theorem\label{th.vanKampen}

1152: Under the assumptions~\loccit{all.but.1}, \loccit{no.boundary}

1153: above, one has

1154: $$

1155: \pi_1(\Sigma_k\sminus(C\cup E))=

1156:  \bigl<\zeta_1,\ldots,\zeta_d\bigm|

1157:  m_i=\id,\ i=1,\ldots,l,\ \ \Gr^k=1\bigr>,

1158: $$

1159: where each \emph{braid relation} $m_i=\id$ should be understood as

1160: a $d$-tuple of relations $\zeta_j=m_i(\zeta_j)$, $j=1,\ldots,d$,

1161: and $\Gr\in\pi_F$ is the element introduced in

1162: Section~\ref{s.monodromy.infty}.

1163: \endtheorem

1164: 

1165: The relation $\Gr^k=1$ in Theorem~\ref{th.vanKampen} is called the

1166: \emph{relation at infinity}. If $k=1$, it coincides with the well

1167: known relation $\Gr=1$ for the group of a plane curve.

1168: 

1169: \proof

1170: First, consider the case when $\disk$ contains \emph{all} singular

1171: fibers of~$C$. As in the proof of

1172: Proposition~\ref{monodromy.infty}, one can replace~$\disk$ with

1173: any larger disk, \eg, with the one given by

1174: $\{\ls|x|\le1/\epsilon\}$, where $(x,y)$ are affine coordinates

1175: such that the point $x=\infty$ is not in~$\disk$ and $\epsilon$ is

1176: a sufficiently small positive real number. Furthermore, one can

1177: take for~$s$ a constant section $x\mapsto\epsilon^{-k}c=\const$,

1178: $\ls|c|\gg0$,

1179: see Remark~\ref{constant.section}, and choose the base point~$b$

1180: in the boundary $\partial\disk$.

1181: The fundamental group

1182: $\pi_1(p^{-1}(\disk)\sminus(C\cup E))$ is given by

1183: Theorem~\ref{th.vanKampen.original}, and the patching of the

1184: nonsingular fiber $\{x=\infty\}=\{x'=0\}$

1185: results in the additional relation $[\partial\Gamma]=1$, where

1186: $\Gamma$ is the disk $\{y'=c,\ \ls|x'|\le\epsilon\}$. (Here,

1187: $x'=1/x$ and $y'=y/x^k$ are the affine coordinates in the

1188: complementary chart, see Remark~\ref{affine}.

1189: We assume that the constant $\ls|c|$ is so large that

1190: $\Gamma\cap\Conv C=\varnothing$.) Restricting to the

1191: boundary $x'=\epsilon\exp(-2\pi t)$, $t\in[0,1]$, and passing back

1192: to $(x,y)$, one finds that the loop $\partial\Gamma$ is given by

1193: $x=\epsilon^{-1}\exp(2\pi t)$, $y=\epsilon^{-k}c\exp(2k\pi t)$;

1194: it is homotopic to $\Gr^k\cdot[s(\partial\disk)]$. Since the loop

1195: $s(\partial\disk)$ is contractible (along the image of~$s$), the

1196: extra relation is $\Gr^k=1$, as stated.

1197: 

1198: Now, assume that one singular fiber of~$C$ is not in~$\disk$. Extend

1199: $\disk$ to a larger disk $\disk'\supset\disk$ containing the

1200: missing singular fiber (and extend the braid monodromy, see

1201: Remark~\ref{extension}). For~$\disk'$, the theorem has already

1202: been proved, and the resulting presentation of the group differs

1203: from the one given by~$\disk$ by an extra relation $m_{l+1}=\id$.

1204: However, under an appropriate choice of the

1205: additional generator~$\Gs_{l+1}$,

1206: one has

1207: $[\partial\disk']=[\partial\disk]\cdot\Gs_{l+1}$. Clearly,

1208: $m([\partial\disk])$ is a word in $m_1,\ldots,m_l$ and, in view of

1209: Proposition~\ref{monodromy.infty}, the monodromy

1210: $m([\partial\disk'])$ is the conjugation by~$\Gr^{-k}$. Hence, in

1211: the presence of the relation at infinity $\Gr^k=1$, the

1212: additional relation

1213: $m_{l+1}=\id$ is a consequence of the other braid relations,

1214: and the statement follows.

1215: \endproof

1216: 

1217: 

1218: \section{The fundamental group\label{S.group}}

1219: 

1220: \subsection{Preliminaries}

1221: Fix a sextic~$B$, pick a stable involutive symmetry~$c$ of~$B$,

1222: see~\S\ref{S.model}, and let $\B,\L\subset\Sigma_2=\Cp2(O_c)/c$ be

1223: the projections of~$B$ and~$L_c$, respectively. We start with

1224: applying Theorem~\ref{th.vanKampen} to the $4$-gonal curve $\B+\L$

1225: and computing the group

1226: $\bpi:=\pi_1(\Sigma_2\sminus(\B\cup\L\cup E))$.

1227: %the

1228: %group $\bpi:=\pi_1(\Sigma_2\sminus(\B\cup E\cup\L))$ and compute

1229: %it,

1230: %applying the classical Zariski--van Kampen method~\cite{vanKampen}

1231: %to the ruling of~$\Sigma_2$ (the pencil $\{x=\const\}$ in the notation

1232: %of~\S\ref{S.model}).

1233: 

1234: In order to visualize the braid monodromy, we

1235: will consider the standard \emph{real structure} (\ie,

1236: anti-holomorphic involution) $\conj\:(x,y)\mapsto(\bar x,\bar y)$

1237: on~$\Sigma_2$, where bar stands for the complex conjugation. A

1238: reduced algebraic

1239: curve $C\subset\Sigma_2$ is said to be \emph{real} (with respect

1240: to~$\conj$)

1241: if it is $\conj$-invariant (as a set).

1242: Alternatively, $C$ is real if

1243: and only if, in

1244: the coordinates $(x,y)$, it can be given by a polynomial with real

1245: coefficients. In particular, the curve~$\B$ given

1246: by~\eqref{eq.equation} is real. Given a real curve

1247: $C\subset\Sigma_2$, one can speak about its \emph{real part}

1248: $C_{\R}$ (\ie, the set of points of~$C$ fixed by~$\conj$), which

1249: is a codimension~$1$ subset in the real part

1250: of~$\Sigma_2$.

1251: 

1252: To use Theorem~\ref{th.vanKampen}, we take for~$\disk$ a closed

1253: regular neighborhood of the smallest segment of the real axis

1254: $\Rp1$ containing all singular fibers of $\B+\L$ except the one of

1255: type~$\tilde\bE_6$ at infinity, see the shaded area in

1256: Figure~\ref{fig.basis}, right.

1257: Recall that

1258: singular are the fiber $\{x=0\}$ through the cusp, the vertical

1259: tangent $\{x=4\}$, and the fibers through the points of

1260: intersection of~$\B$ and~$\L$. (As in~\S\ref{S.model}, we use the

1261: value $r=3$ for the numeric evaluation.)

1262: We only consider the

1263: four extremal sections~$\L$ given by~\eqref{eq.2e6+2a2+a3},

1264: \eqref{eq.3e6+a1}, \eqref{eq.2e6+a5+a2.1},

1265: and~\eqref{eq.2e6+a5+a2.2}. In each case, all

1266: singular fibers are real; they are listed in~\S\ref{S.model}.

1267: 

1268: To compute the braid monodromy, we use a constant real section

1269: $s\:\disk\to\Sigma_2$

1270: given by $x\mapsto\const\gg0$, see Remark~\ref{constant.section}, and

1271: the base point $b=(\epsilon,0)\in\disk$, where $\epsilon>0$ is

1272: sufficiently small.

1273: The basis $\Gs_1,\ldots,\Gs_l$ for the group

1274: $\pi_1(\disk^\sharp,b)$ is chosen as shown in

1275: Figure~\ref{fig.basis}, right: each~$\Gs_i$ is a small loop about

1276: a singular fiber connected to~$b$ by a real segment,

1277: circumventing the

1278: interfering singular fibers in the counterclockwise direction. Let

1279: $F=F_b$ be the base fiber, and choose a basis $\Ga$, $\Gb$, $\Gg$,

1280: $\Gd$ for the group $\pi_F=\pi_1(F^\circ\sminus(\B\cup\L),s(b))$

1281: as shown in Figure~\ref{fig.basis}, left. (Note that, in all cases

1282: considered below, all points of the intersection $F\cap(\B\cup\L)$

1283: are real.)

1284: The following notation

1285: convention is important for the sequel.

1286: 

1287: \Remark\label{notation}

1288: We use a double notation for the elements of the basis

1289: for~$\pi_F$. On the one hand, to be consistent with

1290: Theorem~\ref{th.vanKampen}, we denote them

1291: $\zeta_1,\ldots,\zeta_4$, numbering the loops consecutively according

1292: to the decreasing of the $y$-coordinate of the point. Then the

1293: element $\Gr\in\pi_F$ introduced in

1294: Section~\ref{s.monodromy.infty} is given by

1295: $\Gr=\zeta_1\zeta_2\zeta_3\zeta_4$, and the relation at infinity in

1296: Theorem~\ref{th.vanKampen} turns to

1297: $(\zeta_1\zeta_2\zeta_3\zeta_4)^2=1$. On the other hand, to make

1298: the formulas more readable, we denote the basis elements by $\Ga$,

1299: $\Gb$, $\Gg$, and~$\Gd$. The first three elements are numbered

1300: consecutively, whereas~$\Gd$ plays a

1301: very special r\^ole in the passage

1302: to the group $\pi_1(\Cp2\sminus B)$, see

1303: Lemma~\ref{bpi->pi} below: we always assume that $\Gd$ is

1304: the element represented

1305: by a loop about the point $F\cap\L$.

1306: Thus, the position of~$\Gd$ in the sequence

1307: $(\Ga,\Gb,\Gg,\Gd)$ may change; this position is important for the

1308: expression for~$\Gr$ and hence for the

1309: relation at infinity.

1310: \endRemark

1311: 

1312: %Pick and fix a real section $S=\{y=\const\gg0\}$ of~$\Sigma_2$ and

1313: %a real nonsingular fiber $F=\{x=\epsilon\}$, where $\epsilon>0$ is

1314: %sufficiently small.

1315: %Let $P=F\cap S$, and pick a basis $\Ga$, $\Gb$, $\Gg$, $\Gd$ for the

1316: %group $\pi_F:=\pi_1(F\sminus(\B\cup E\cup\L),P)$ as shown in

1317: %Figure~\ref{fig.basis}, left. (Note that, in all cases considered

1318: %below, all intersection points are real.) We always assume that

1319: %$\Ga$, $\Gb$, $\Gg$ are small loops about the three points of

1320: %$F\cap\B$, numbered consecutively, whereas $\Gd$ is a loop about

1321: %$F\cap\L$. Thus, the position of~$\Gd$ in the sequence

1322: %$(\Ga,\Gb,\Gg,\Gd)$ may change; this position is important for the

1323: %relation at infinity, see below.

1324: 

1325: \midinsert

1326: \centerline{\picture{basis}}

1327: \figure\label{fig.basis}

1328: The basis $\Ga$, $\Gb$, $\Gg$, $\Gd$ and the loops~$\Gs_i$

1329: \endfigure

1330: \endinsert

1331: 

1332: %Let $\Gs_1,\ldots,\Gs_k$ be the basis for the group

1333: %$\pi_1(S\sminus(\bigcup F_i\cup\{x=\infty\}))$ shown in

1334: %Figure~\ref{fig.basis}, right: each~$\Gs_i$ is a small loop about

1335: %$S\cap F_i$ connected to~$P$ by a real segment, circumventing the

1336: %interfering fibers in the counterclockwise direction. For

1337: %each~$i$, let $m_i\:\pi_F\to\pi_F$ be the \emph{braid monodromy}

1338: %along~$\Gs_i$, \ie, the automorphism of~$\pi_F$ obtained by

1339: %dragging~$F$ along~$\Gs_i$ while keeping the base point

1340: %on~$\Gs_i$. Then, the Zariski--van Kampen theorem~\cite{vanKampen}

1341: %states that

1342: %$$

1343: %\bpi=\bigl<\Ga,\Gb,\Gg,\Gd\bigm|

1344: % \text{$m_i=\id$, $i=1,\ldots,k$, $(\Ga\Gb\Gg\Gd)^2=1$}\bigr>.

1345: %\eqtag\label{eq.vanKampen}

1346: %$$

1347: %Here, each \emph{braid relation} $m_i=\id$ should be understood as

1348: %a quadruple of relations $m_i(\Ga)=\Ga$, $m_i(\Gb)=\Gb$,

1349: %$m_i(\Gg)=\Gg$, $m_i(\Gd)=\Gd$; the precise form of the

1350: %\emph{relation at infinity} $(\Ga\Gb\Gg\Gd)^2=1$ depends on the

1351: %order of the generators.

1352: 

1353: %\Remark

1354: %The\mnote{new}

1355: %relation at infinity is an analog of classical van Kampen's

1356: %relation at infinity in the case of $\Cp2=\Sigma_1/E$; it results

1357: %from patching the fiber of~$\Sigma_2$ over $x=\infty$ (which

1358: %should be removed in order to fix a trivialization of the ruling

1359: %of~$\Sigma_2$ and hence define the braid monodromy.)

1360: %The relation

1361: %has the form $[\gamma]=1$, where $\gamma$ is a large circle

1362: %in~$S$ encompassing all points of the intersection

1363: %$S\cap F_1$ and $S\cap B$.

1364: %If $S$ is a small perturbation of $2F'+E$,

1365: %where $F'$ is a fiber close to~$F$ and $F$ is `close to infinity',

1366: %one obviously has $\gamma=(\Ga\Gb\Gg\Gd)^2$.

1367: %In general, the relation has

1368: %the same form due to the fact that

1369: %the product $\Ga\Gb\Gg\Gd$ (in the appropriate

1370: %order) is invariant under the braid group action. For details,

1371: %see~\cite{degt.kplets}.

1372: %\endRemark

1373: 

1374: %\Remark

1375: %As shown in~\cite{degt.Oka3}, in the\mnote{new}

1376: %presence of the relation at infinity, (any) one of the braid relations

1377: %can be ignored. In what follows, we systematically ignore

1378: %the monodromy about the type~$\tilde\bE_6$ singular fiber of~$\B$.

1379: %Then, the computation of the braid monodromy is concentrated in a

1380: %large disk containing the other singular fibers. Hence,

1381: %instead of removing the type~$\tilde\bE_6$ fiber and a

1382: %separate fiber at infinity,

1383: %%in order to simplify the defining equations,

1384: %we can place the type~$\bE_6$

1385: %singular point of~$\B$ over $x=\infty$

1386: %(this simplifies the defining equation of the curve),

1387: %merely adding the standard relation at

1388: %infinity to the remaining braid relations.

1389: %\endRemark

1390: 

1391: The passage from a presentation of~$\bpi$ to the

1392: that of the group $\pi_1:=\pi_1(\Cp2\sminus B)$ is

1393: given by the following lemma.

1394: 

1395: \lemma\label{bpi->pi}

1396: If $\bpi$ is given by

1397: $\bigl<\Ga,\Gb,\Gg,\Gd\bigm|R_j=1,\ j=1,\ldots,s\bigr>$, then

1398: $$

1399: \pi_1=\bigl<\Ga,\bGa,\Gb,\bGb,\Gg,\bGg\bigm|R'_j=\bar R'_j=1,\

1400:  j=1,\ldots,s\bigr>,

1401: $$

1402: where bar stands for the conjugation by~$\Gd$,

1403: $\bar w=\Gd^{-1}w\Gd$,

1404: each relation $R'_j$ is obtained from $R_j$, $j=1,\ldots,s$,

1405: by letting

1406: $\Gd^2=1$ and expressing the result in terms of the generators

1407: $\Ga$, $\bGa$, \dots, and $\bar R'_j=\Gd^{-1}R'_j\Gd$,

1408: $j=1,\ldots,s$.

1409: \rom(In other words,

1410: $\bar R'_j$ is obtained from $R'_j$ by interchanging

1411: $\Ga\leftrightarrow\bGa$, $\Gb\leftrightarrow\bGb$, and

1412: $\Gg\leftrightarrow\bGg$.\rom)

1413: \endlemma

1414: 

1415: \proof

1416: The projection

1417: $\Cp2\sminus(B\cup O_c)\to\Sigma_2\sminus(\B\cup E)$ is a double

1418: covering ramified at~$\L$. Hence, one has

1419: $\pi_1=\pi_1(\Cp2\sminus(B\cup O_c))=\Ker[\kappa\:\bpi/\Gd^2\to\CG2]$, where

1420: $\kappa\:\Ga,\Gb,\Gg\mapsto0$ and $\kappa\:\Gd\mapsto1$. (Note

1421: that the compactification of the

1422: double covering above is \emph{not} ramified at~$\B$.) Lift

1423: $\kappa$ to a homomorphism

1424: $\tilde\kappa\:\<\Ga,\Gb,\Gg,\Gd>\to\Z_2$. The two cosets modulo

1425: $\Ker\tilde\kappa$ are represented by~$1$ and~$\Gd$, and the

1426: standard calculation shows that $\Ker\tilde\kappa$ is the free

1427: group generated by $\Ga,\bGa,\Gb,\bGb,\Gg,\bGg,\Gd^2$. The

1428: kernel~$N$ of the epimorphism $\Ker\tilde\kappa\onto\pi_1$ is

1429: normally generated \emph{in $\<\Ga,\Gb,\Gg,\Gd>$} by~$\Gd^2$ and

1430: $R'_j$, $j=1,\ldots,s$. Hence, one can remove

1431: the generator~\smash{$\Gd^2$} from the

1432: presentation. Besides,

1433: since the conjugation by~$\Gd$ is not an

1434: inner automorphism of $\Ker\tilde\kappa$, one should add the

1435: conjugates $\bar R'_j=\Gd^{-1}R'_j\Gd$ to obtain a set normally

1436: generating~$N$ in $\Ker\tilde\kappa$. The resulting presentation

1437: of~$\pi_1$ is the one stated in the lemma.

1438: \endproof

1439: 

1440: \Remark

1441: Note that $\spbar\:w\mapsto\bar w=\Gd w\Gd$ is an involutive

1442: automorphism of~$\pi_1$. Hence, whenever a relation $R=1$ holds

1443: in~$\pi_1$, the relation $\bar R=1$ also holds.

1444: \endRemark

1445: 

1446: \subsection{The set of singularities $(3\bE_6)\splus\bA_1$}

1447: \label{s.3e6+a1}

1448: Take for~$\L$ the section given by~\eqref{eq.3e6+a1}. The pair

1449: $(\B,\L)$ looks as shown in Figure~\ref{fig.2e6+2a2+a3}, and the

1450: singular fibers are listed in~\ref{s.cusp}.

1451: The generators

1452: $\zeta_1=\Ga$, $\zeta_2=\Gd$, $\zeta_3=\Gb$, $\zeta_4=\Gg$

1453: for~$\bpi$ are subject

1454: to the relations

1455: $$

1456: \alignat2

1457: &(\Gd\Gb)^2=(\Gb\Gd)^2&\qquad&

1458:  \text{(the tangency point $x=2$)},\\\allowdisplaybreak

1459: &(\Gd\Gb)\Gb(\Gd\Gb)\1=\Gg&&

1460:  \text{(the vertical tangent $x=4$)},\\\allowdisplaybreak

1461: &[\Gd,\Ga\Gd\Gb\Ga]=1,\quad%&&\text{(the cusp)},\\

1462: \Ga\Gd\Gb\Ga=\Gb\Ga\Gd\Gb&&

1463:  \text{(the cusp $x=0$)},\\\allowdisplaybreak

1464: &[\Gd,(\Ga\Gd\Gb)\Gg(\Ga\Gd\Gb)\1]=1&&

1465:  \text{(the transversal intersection $x=-16$)},\\\allowdisplaybreak

1466: &(\Ga\Gd\Gb\Gg)^2=1&&\text{(the relation at infinity)}.

1467: \endalignat

1468: $$

1469: Letting $\Gd^2=1$ and passing to $\Ga$, $\bGa$, $\Gb$, $\bGb$,

1470: $\Gg$, $\bGg$, see Lemma~\ref{bpi->pi}, one can rewrite these

1471: relations in the following form:

1472: $$

1473: \gather

1474: [\Gb,\bGb]=1,

1475: \eqtag\label{eq.1.1}\\\allowdisplaybreak

1476: \Gg=\bGb,\quad \bGg=\Gb,

1477: \eqtag\label{eq.1.2}\\\allowdisplaybreak

1478: \Ga\bGb\bGa=\bGa\Gb\Ga=\Gb\Ga\bGb=\bGb\bGa\Gb,

1479: \eqtag\label{eq.1.3}\\\allowdisplaybreak

1480: \Ga\Gb\Ga\1=\bGa\bGb\bGa\1,

1481: \eqtag\label{eq.1.4}\\\allowdisplaybreak

1482: \Ga\bGb\Gb\bGa\Gb\bGb=1.

1483: \eqtag\label{eq.1.5}

1484: \endgather

1485: $$

1486: (In~\eqref{eq.1.4} and~\eqref{eq.1.5}, we eliminate~$\Gg$

1487: using~\eqref{eq.1.2}.) Now, one can use the last

1488: relation in~\eqref{eq.1.3} to eliminate~$\bGa$: one has

1489: $\bGa=\bGb\1\Gb\Ga\bGb\Gb\1$. Substituting this expression to

1490: $\Ga\bGb\bGa=\Gb\Ga\bGb$ and $\bGa\Gb\Ga=\Gb\Ga\bGb$

1491: in~\eqref{eq.1.3} and

1492: using~\eqref{eq.1.1}, one obtains, respectively, the braid

1493: relations $\Ga\Gb\Ga=\Gb\Ga\Gb$ and $\Ga\bGb\Ga=\bGb\Ga\bGb$.

1494: Conjugating by~$\Gd$, one also has $\bGa\Gb\bGa=\Gb\bGa\Gb$

1495: and $\bGa\bGb\bGa=\bGb\bGa\bGb$. Then, \eqref{eq.1.4} turns to

1496: $\Gb\1\Ga\Gb=\bGb\1\bGa\bGb$ and, eliminating~$\bGa$, one obtains

1497: $[\Ga,\bGb^2\Gb\2]=1$. Finally, eliminating~$\bGa$ from the last

1498: relation~\eqref{eq.1.5}, one gets $\Ga\Gb^2\Ga\bGb^2=1$. Thus, the

1499: map $\Gb\mapsto\Gs_1$, $\Ga\mapsto\Gs_2$, $\bGb\mapsto\Gs_3$

1500: establishes an isomorphism

1501: $$

1502: \pi_1(\Cp2\sminus B)=\BG4/

1503:  \<[\Gs_2,\Gs_1^2\Gs_3\2],\ \Gs_2\Gs_1^2\Gs_2\Gs_3^2>.

1504: $$

1505: It remains to notice that, in the presence of the second relation

1506: in the presentation above, the first one turns into

1507: $[\Gs_2,\Gs_1^2\Gs_2\Gs_1^2\Gs_2]=1$, or

1508: $[\Gs_2,(\Gs_1\Gs_2)^3]=1$, which holds automatically.

1509: Thus, one has

1510: $$

1511: \pi_1(\Cp2\sminus B)=\BG4/\Gs_2\Gs_1^2\Gs_2\Gs_3^2.

1512: \eqtag\label{eq.G0}

1513: $$

1514: 

1515: \corollary\label{e6->>pi}

1516: Let~$\MB$ be

1517: a Milnor ball about a type~$\bE_6$ singular point of~$B$.

1518: Then the inclusion homomorphism

1519: $\pi_1(\MB\sminus B)\to\pi_1(\Cp2\sminus B)$ is onto.

1520: \endcorollary

1521: 

1522: \proof

1523: Since any pair of type~$\bE_6$ singular points can be permuted by

1524: a stable symmetry of~$B$, see~\cite{symmetric},

1525: it suffices to prove the

1526: statement for the type~$\bE_6$ point

1527: resulting from the cusp of~$\B$. In

1528: this case, the statement follows from~\eqref{eq.1.2}, as $\Ga$,

1529: $\bGa$, $\Gb$, and~$\bGb$ are all

1530: in the image of $\pi_1(\MB\sminus B)$.

1531: \endproof

1532: 

1533: \subsection{The set of singularities $(2\bE_6\splus2\bA_2)\splus\bA_3$}

1534: \label{s.2e6+2a2+a3}

1535: Take for~$\L$ the section given by~\eqref{eq.2e6+2a2+a3}. The pair

1536: $(\B,\L)$ is plotted in Figure~\ref{fig.2e6+2a2+a3}, and the

1537: singular fibers are listed in~\ref{s.tangent}.

1538: The generators

1539: $\zeta_1=\Gd$, $\zeta_2=\Ga$, $\zeta_3=\Gb$, $\zeta_4=\Gg$

1540: for~$\bpi$ are subject

1541: to the relations

1542: %We choose the generators $\Ga$, $\Gd$, $\Gb$, $\Gg$ in a

1543: %nonsingular fiber between the cusp and the tangency point. The

1544: %relations are:

1545: $$

1546: \alignat2

1547: &[\Gd,\Ga]=1&\qquad&

1548:  \text{(the transversal intersection $x\approx.0459$)},\\\allowdisplaybreak

1549: &\Ga\Gb\Ga=\Gb\Ga\Gb&&

1550:  \text{(the cusp $x=0$)},\\\allowdisplaybreak

1551: &[\Gd,\Gb\Ga\1\Gg\Ga\Gb\1]=1&&

1552:  \text{(the transversal intersection $x\approx-19.1$)},\\\allowdisplaybreak

1553: &(\Gd\Gb)^4=(\Gb\Gd)^4&&

1554:  \text{(the tangency point $x=.625$)},\\\allowdisplaybreak

1555: &(\Gd\Gb)^2\Gb(\Gd\Gb)\2=\Gg&&

1556:  \text{(the vertical tangent $x=4$)},\\\allowdisplaybreak

1557: &(\Gd\Ga\Gb\Gg)^2=1&&\text{(the relation at infinity)}.

1558: \endalignat

1559: $$

1560: (The third relation is simplified using $[\Gd,\Ga]=1$.)

1561: Letting $\Gd^2=1$ and passing to $\Ga=\bGa$,

1562: $\Gb$, $\bGb$, $\Gg$, $\bGg$, see Lemma~\ref{bpi->pi},

1563: one can rewrite these

1564: relations as follows:

1565: $$

1566: \gather

1567: \Ga=\bGa,

1568: \eqtag\label{eq.2.1}\\\allowdisplaybreak

1569: \Ga\Gb\Ga=\Gb\Ga\Gb,\quad \Ga\bGb\Ga=\bGb\Ga\bGb,

1570: \eqtag\label{eq.2.2}\\\allowdisplaybreak

1571: \Gb\Ga\1\bGb\Gb\bGb\1\Ga\Gb\1=\bGb\Ga\1\Gb\bGb\Gb\1\Ga\bGb\1,

1572: \eqtag\label{eq.2.3}\\\allowdisplaybreak

1573: (\bGb\Gb)^2=(\Gb\bGb)^2,

1574: \eqtag\label{eq.2.4}\\\allowdisplaybreak

1575: \bGb\Gb\bGb\1=\Gg,\quad\Gb\bGb\Gb\1=\bGg,

1576: \eqtag\label{eq.2.5}\\\allowdisplaybreak

1577: \Ga\bGb\Gb\bGb\Gb\1\Ga\Gb\bGb\Gb\bGb\1=1.

1578: \eqtag\label{eq.2.6}

1579: \endgather

1580: $$

1581: (We use~\eqref{eq.2.1} and~\eqref{eq.2.5} to eliminate~$\bGa$,

1582: $\Gg$, and~$\bGg$ in the other relations.) Thus,

1583: $$

1584: \pi_1(\Cp2\sminus B)=G_3:=\bigl<\Ga,\Gb,\bGb\bigm|

1585:  \text{\eqref{eq.2.2}--\eqref{eq.2.4}, \eqref{eq.2.6}}\bigr>.

1586: \eqtag\label{eq.G3}

1587: $$

1588: 

1589: \midinsert

1590: \centerline{\picture{basis2}}

1591: \figure

1592: Generators in $F'=\{x=b'=\const\ll0\}$

1593: \endfigure\label{fig.basis2}

1594: \endinsert

1595: 

1596: The following statement is a consequence of the

1597: monodromy computation.

1598: 

1599: \lemma\label{e6.a3}

1600: Let $\Ga_1$, $\Gb_1$, $\Gg_1$, $\Gd_1$ be the basis in a fiber

1601: $F'=\{x=\const\ll0\}$ shown in Figure~\ref{fig.basis2}, left.

1602: Then,

1603: considering $\Ga_1$, $\Gb_1$, and $\Gg_1$ as elements of~$\bpi$,

1604: one has

1605: $\Ga_1=\bGb$, $\Gb_1=\Gb\1\Ga\Gb$, and $\Gg_1=\Gg$.

1606: \qed

1607: \endlemma

1608: 

1609: \corollary\label{e6->>pi.a3}

1610: Let~$\MB$ be

1611: a Milnor ball about a type~$\bE_6$ singular point of~$B$.

1612: Then the inclusion homomorphism

1613: $\pi_1(\MB\sminus B)\to\pi_1(\Cp2\sminus B)$ is onto.

1614: \endcorollary

1615: 

1616: \proof

1617: In view of~\eqref{eq.2.5}, one has $\Gb=\Ga_1\1\Gg_1\Ga_1$.

1618: Then $\Ga=\Gb\Gb_1\Gb\1$; hence, the elements

1619: $\Ga_1$, $\Gb_1$, and~$\Gg_1$ generate the group. On the other

1620: hand, $\Ga_1$, $\Gb_1$, $\Gg_1$ are in the image of

1621: $\pi_1(\MB\sminus B)$.

1622: \endproof

1623: 

1624: \subsection{The set of singularities

1625: $(2\bE_6\splus\bA_5)\splus\bA_2$:

1626: %the family~\eqref{eq.2e6+a5+a2.1}}

1627: the first family}

1628: \label{s.2e6+a5+a2.1}

1629: Take for $\L$ the section given by~\eqref{eq.2e6+a5+a2.1}.

1630: The pair $(\B,\L)$ looks as shown in

1631: Figure~\ref{fig.2e6+2a2+a3}, and the singular fibers are listed

1632: in~\ref{s.inflection}.

1633: The generators

1634: $\zeta_1=\Ga$, $\zeta_2=\Gb$, $\zeta_3=\Gd$, $\zeta_4=\Gg$

1635: satisfy the following

1636: relations:

1637: $$

1638: \alignat2

1639: &[\Gd,\Ga\Gb]=1,\quad \Gd\Ga\Gb\Ga=\Gb\Ga\Gb\Gd

1640:  &&\text{(the cusp $x=0$)},\\\allowdisplaybreak

1641: &(\Gb\Gd)^3=(\Gd\Gb)^3&&

1642:  \text{(the tangency point $x\approx.951$)},\\\allowdisplaybreak

1643: &(\Gb\Gd)\Gb(\Gb\Gd)\1=\Gg&&

1644:  \text{(the vertical tangent $x=4$)},\\\allowdisplaybreak

1645: &[\Gd,\Ga\1\Gg\Ga]=1&&

1646:  \text{(the transversal intersection $x\approx-18.4$)},\\\allowdisplaybreak

1647: &(\Ga\Gb\Gd\Gg)^2=1&\qquad&\text{(the relation at infinity)}.

1648: \endalignat

1649: $$

1650: Letting $\Gd^2=1$ and passing to $\Ga$,

1651: $\bGa$, $\Gb$, $\bGb$, $\Gg$, $\bGg$, see Lemma~\ref{bpi->pi},

1652: one obtains

1653: $$

1654: \gather

1655: \Ga\Gb=\bGa\bGb,\quad\bGa\bGb\bGa=\Gb\Ga\Gb,\quad

1656:  \Ga\Gb\Ga=\bGb\bGa\bGb,

1657: \eqtag\label{eq.3.1}\\\allowdisplaybreak

1658: \bGb\Gb\bGb=\Gb\bGb\Gb,

1659: \eqtag\label{eq.3.2}\\\allowdisplaybreak

1660: \Gb\bGb\Gb\1=\Gg,\quad\bGb\Gb\bGb\1=\bGg,

1661: \eqtag\label{eq.3.3}\\\allowdisplaybreak

1662: \Ga\1\Gg\Ga=\bGa\1\bGg\bGa,

1663: \eqtag\label{eq.3.4}\\\allowdisplaybreak

1664: \Ga\Gb\bGg\Ga\Gb\Gg=1.

1665: \eqtag\label{eq.3.5}

1666: \endgather

1667: $$

1668: The cusp relations~\eqref{eq.3.1} can be rewritten in the form

1669: $$

1670: \bGa=(\Ga\Gb)\1\Gb(\Ga\Gb),\quad

1671: \bGb=(\Ga\Gb)\Ga(\Ga\Gb)\1,\quad

1672: (\Ga\Gb)^3=(\Gb\Ga)^3,

1673: \eqtag\label{eq.3.6}

1674: $$

1675: or, in terms of~$\bGa$, $\bGb$, in the form

1676: $$

1677: \Ga=(\bGa\bGb)\1\bGb(\bGa\bGb),\quad

1678: \Gb=(\bGa\bGb)\bGa(\bGa\bGb)\1,\quad

1679: (\bGa\bGb)^3=(\bGb\bGa)^3.

1680: \eqtag\label{eq.3.7}

1681: $$

1682: Geometrically, one has

1683: $\pi_1(\MB\sminus B)=\bigl<\Ga,\Gb\bigm|(\Ga\Gb)^3=(\Gb\Ga)^3\bigr>$,

1684: where $\MB$ is a Milnor ball around the type~$\bA_5$ singular

1685: point.

1686: 

1687: Writing~\eqref{eq.3.5} as $\Ga\Gb\bGg\bGa\bGb\Gg=1$ and

1688: eliminating~$\Gg$ and~$\bGg$ using~\eqref{eq.3.3}

1689: and~\eqref{eq.3.2}, we can rewrite this relation in the form

1690: $$

1691: \Ga\bGb\Gb\bGa\Gb\bGb=1.

1692: \eqtag\label{eq.3.8}

1693: $$

1694: Eliminating~$\Gg$ and~$\bGg$ from~\eqref{eq.3.4}, we obtain

1695: $$

1696: \Ga\1\Gb\bGb\Gb\1\Ga=\bGa\1\bGb\Gb\bGb\1\bGa.

1697: \eqtag\label{eq.3.9}

1698: $$

1699: Thus, we have

1700: $$

1701: \pi_1(\Cp2\sminus B)=G_2':=\bigl<\Ga,\Gb\bigm|

1702:  \text{$(\Ga\Gb)^3=(\Gb\Ga)^3$, \eqref{eq.3.2}, \eqref{eq.3.8},

1703:  \eqref{eq.3.9}}\bigr>,

1704: \eqtag\label{eq.G2.1}

1705: $$

1706: where $\bGa$ and $\bGb$ are the words given by~\eqref{eq.3.6}. I

1707: could not find any substantial simplification of this

1708: presentation. An alternative presentation of~$G_2'$ (as well as

1709: of the group~$G_2''$ introduced in~\eqref{eq.G2.2} below)

1710: is given in Eyral, Oka~\cite{EyralOka}.

1711: 

1712: As a part of computing the braid monodromy, we get the following

1713: lemma.

1714: 

1715: \lemma\label{e6.a5.1}

1716: Let $\Ga_1$, $\Gb_1$, $\Gg_1$, $\Gd_1$ be the basis in a fiber

1717: $F'=\{x=\const\ll0\}$ shown in Figure~\ref{fig.basis2}, left.

1718: Then,

1719: considering $\Ga_1$, $\Gb_1$, and $\Gg_1$ as elements of~$\bpi$,

1720: one has

1721: $\Ga_1=\Gb$, $\Gb_1=\bGb\1\bGa\bGb$, and $\Gg_1=\Gg$.

1722: \qed

1723: \endlemma

1724: 

1725: \corollary\label{e6->>pi.a5.1}

1726: Let~$\MB$ be

1727: a Milnor ball about a type~$\bE_6$ singular point of~$B$.

1728: Then the inclusion homomorphism

1729: $\pi_1(\MB\sminus B)\to\pi_1(\Cp2\sminus B)$ is onto.

1730: \endcorollary

1731: 

1732: \proof

1733: Due to~\eqref{eq.3.3}, one has $\bGb=\Ga_1\1\Gg_1\Ga_1$. Then

1734: $\bGa=\bGb\Gb_1\bGb\1$ and, in view of~\eqref{eq.3.7}

1735: and~\eqref{eq.3.3}, $\bGa$ and $\bGb$ generate the group.

1736: \endproof

1737: 

1738: 

1739: \subsection{The set of singularities

1740: $(2\bE_6\splus\bA_5)\splus\bA_2$:

1741: %the family~\eqref{eq.2e6+a5+a2.2}}

1742: the second family}

1743: \label{s.2e6+a5+a2.2}

1744: Now, let~$\L$ be the section given by~\eqref{eq.2e6+a5+a2.2}. The

1745: pair $(\B,\L)$ is plotted in Figure~\ref{fig.2e6+a5+a2}, and the

1746: singular fibers are listed in~\ref{s.inflection}.

1747: The generators for~$\pi_F$ are

1748: $\zeta_1=\Ga$, $\zeta_2=\Gb$, $\zeta_3=\Gd$, $\zeta_4=\Gg$,

1749: and the

1750: relations are:

1751: $$

1752: \alignat2

1753: &[\Gd,\Ga\Gb]=1,\quad \Gd\Ga\Gb\Ga=\Gb\Ga\Gb\Gd

1754:  &&\text{(the cusp $x=0$)},\\\allowdisplaybreak

1755: &(\Gg\Gd)^3=(\Gd\Gg)^3&&

1756:  \text{(the tangency point $x\approx3.55$)},\\\allowdisplaybreak

1757: &(\Gd\Gg\Gd)\Gg(\Gd\Gg\Gd)\1=\Gb&&

1758:  \text{(the vertical tangent $x=4$)},\\\allowdisplaybreak

1759: &[\Gd,\Gg\Ga\Gg\1]=1&&

1760:  \text{(the transversal intersection $x\approx4.94$)},\\\allowdisplaybreak

1761: &(\Ga\Gb\Gd\Gg)^2=1&\qquad&\text{(the relation at infinity)}.

1762: \endalignat

1763: $$

1764: Let $\Gd^2=1$ and pass to the generators $\Ga$,

1765: $\bGa$, $\Gb$, $\bGb$, $\Gg$, $\bGg$, see Lemma~\ref{bpi->pi}.

1766: Then,

1767: in addition to the cusp relations~\eqref{eq.3.6}

1768: (or~\eqref{eq.3.1}\,) and relation at infinity~\eqref{eq.3.5},

1769: we obtain

1770: $$

1771: \gather

1772: \Gg\bGg\Gg=\bGg\Gg\bGg,

1773: \eqtag\label{eq.4.1}\\\allowdisplaybreak

1774: \bGg\Gg\bGg\1=\Gb,\quad\Gg\bGg\Gg\1=\bGb

1775: \eqtag\label{eq.4.2}\\\allowdisplaybreak

1776: \Gg\Ga\Gg\1=\bGg\bGa\bGg\1.

1777: \eqtag\label{eq.4.3}

1778: \endgather

1779: $$

1780: Thus,

1781: $$

1782: \pi_1(\Cp2\sminus B)=G_2'':=\bigl<\Ga,\Gb,\Gg,\bGg\bigm|

1783:  \text{$(\Ga\Gb)^3=(\Gb\Ga)^3$, \eqref{eq.3.5},

1784:  \eqref{eq.4.1}--\eqref{eq.4.3}}\bigr>,

1785: \eqtag\label{eq.G2.2}

1786: $$

1787: where $\bGa$ and $\bGb$ are the words given by~\eqref{eq.3.6}.

1788: Note that one can eliminate either~$\bGg$, using~\eqref{eq.3.5},

1789: or~$\Gb$, using~\eqref{eq.4.2}.

1790: 

1791: Extending the braid monodromy beyond the cusp of~$B$ (to the

1792: negative values of~$x$), we obtain the

1793: following statement.

1794: 

1795: \lemma\label{e6.a5.2}

1796: Let $\Gd_1$, $\Ga_1$, $\Gb_1$, $\Gg_1$ be the basis in a fiber

1797: $F'=\{x=\const\ll0\}$ shown in Figure~\ref{fig.basis2}, right.

1798: Then,

1799: considering $\Ga_1$, $\Gb_1$, and $\Gg_1$ as elements of~$\bpi$,

1800: one has

1801: $\Ga_1=\bGb$, $\Gb_1=\bGb\1\bGa\bGb$, and $\Gg_1=\Gg$.

1802: \qed

1803: \endlemma

1804: 

1805: \corollary\label{e6->>pi.a5.2}

1806: Let~$\MB$ be

1807: a Milnor ball about a type~$\bE_6$ singular point of~$B$.

1808: Then the inclusion homomorphism

1809: $\pi_1(\MB\sminus B)\to\pi_1(\Cp2\sminus B)$ is onto.

1810: \endcorollary

1811: 

1812: \proof

1813: In view of~\eqref{eq.3.7} and~\eqref{eq.3.5}, the elements

1814: $\bGa=\Ga_1\Gb_1\Ga_1\1$, $\bGb=\Ga_1$,

1815: and $\Gg=\Gg_1$ generate the group.

1816: \endproof

1817: 

1818: 

1819: \subsection{Comparing the two groups}

1820: Let~$B'$ and~$B''$ be the sextics considered

1821: in~\ref{s.2e6+a5+a2.1} and~\ref{s.2e6+a5+a2.2}, respectively, so

1822: that their fundamental groups are~$G_2'$ and~$G_2''$. As explained

1823: in Eyral, Oka~\cite{EyralOka}, the profinite completions of~$G_2'$

1824: and~$G_2''$ are isomorphic (as the two curves are conjugate over

1825: an algebraic number field). Whether $G_2'$ and~$G_2''$ themselves

1826: are isomorphic is still an open question. Below, we suggest an

1827: attempt to distinguish the two groups geometrically.

1828: 

1829: \proposition\label{a5->>pi}

1830: Let~$\MB$ be

1831: a Milnor ball about the type~$\bA_5$ singular point of~$B'$.

1832: Then the inclusion homomorphism

1833: $\pi_1(\MB\sminus B')\to\pi_1(\Cp2\sminus B')$ is onto.

1834: \endproposition

1835: 

1836: \proof

1837: According to~\eqref{eq.G2.1}, the group

1838: $\pi_1(\Cp2\sminus B')=G_2'$ is generated by $\Ga$ and~$\Gb$,

1839: which are both in the image of $\pi_1(\MB\sminus B')$.

1840: \endproof

1841: 

1842: \conjecture\label{a5not->>pi}

1843: Let~$\MB$ be

1844: a Milnor ball about the type~$\bA_5$ singular point of~$B''$.

1845: Then the image of the inclusion homomorphism

1846: $\pi_1(\MB\sminus B'')\to\pi_1(\Cp2\sminus B'')$ does not

1847: contain~$\Gg$ or~$\bGg$.

1848: \endconjecture

1849: 

1850: \Remark

1851: If true, Conjecture~\ref{a5not->>pi} together with

1852: Proposition~\ref{a5->>pi} would provide a topological distinction

1853: between pairs $(\Cp2,B')$ and $(\Cp2,B'')$. Note that,

1854: according to~\cite{JAG}, the two pairs are not diffeomorphic.

1855: \endRemark

1856: 

1857: 

1858: \subsection{Other symmetric sets of singularities}

1859: \label{s.stable}

1860: The set of singularities $(3\bE_6)$ is obtained by perturbing~$\L$

1861: in Section~\ref{s.3e6+a1} to a section tangent to~$\B$ at the cusp

1862: and transversal to~$\B$ otherwise. This procedure

1863: replaces~\eqref{eq.1.1} with $\bGb=\Gb$ or, alternatively,

1864: introduces a relation $\Gs_3=\Gs_1$ in~\eqref{eq.G0}. The

1865: resulting group is $\BG3/(\Gs_1\Gs_2)^3$.

1866: 

1867: The sets of singularities of the form

1868: $(2\bE_6\splus2\bA_2)\splus{\ldots}$ are obtained by

1869: perturbing $\L$ in Section~\ref{s.2e6+2a2+a3}.

1870: If $\L$ is perturbed to a double tangent (the set of singularities

1871: $(2\bE_6\splus2\bA_2)\splus2\bA_1$), relation~\eqref{eq.2.4}

1872: is replaced with $[\Gb,\bGb]=1$. Then, \eqref{eq.2.6} turns to

1873: $\Ga\bGb^2\Ga\Gb^2=1$, and \eqref{eq.2.3} turns to

1874: $$

1875: \Gb\underline{\Ga\1\Gb\Ga}\Gb\1=\bGb\underline{\Ga\1\bGb\Ga}\bGb\1.

1876: $$

1877: Replacing the underlined expressions using the braid

1878: relations~\eqref{eq.2.2} converts this relation to

1879: $\Gb^2\Ga\Gb\2=\bGb^2\Ga\bGb\2$, \ie, $[\Ga,\bGb^2\Gb\2]=1$.

1880: As explained in~\ref{s.3e6+a1},

1881: %Hence,

1882: the map

1883: $\Gb\mapsto\Gs_1$, $\Ga\mapsto\Gs_2$, $\bGb\mapsto\Gs_3$

1884: establishes an isomorphism

1885: $\pi_1(\Cp2\sminus B)=\BG4/\Gs_2\Gs_1^2\Gs_2\Gs_3^2$.

1886: 

1887: Any other perturbation of~$\L$ produces an extra point of

1888: transversal intersection with~$\B$, replacing~\eqref{eq.2.4} with

1889: $\Gb=\bGb$. The resulting group is $\BG3/(\Gs_1\Gs_2)^3$.

1890: 

1891: Finally, the sets of singularities $(2\bE_6\splus\bA_5)\splus\bA_1$

1892: and $(2\bE_6\splus\bA_5)$ are obtained by perturbing the inflection

1893: tangency point of~$\L$ and~$\B$ in Section~\ref{s.2e6+a5+a2.1}.

1894: This procedure replaces~\eqref{eq.3.2} with $\bGb=\Gb$. Then, from

1895: the first relation in~\eqref{eq.3.1} one has $\bGa=\Ga$,

1896: relation~\eqref{eq.3.3} results in $\Gg=\bGb=\Gb$,

1897: and relation~\eqref{eq.3.5} turns to $(\Ga\Gb^2)^2=1$.

1898: Hence, the group is $\BG3/(\Gs_1\Gs_2)^3$. (Note that

1899: $(\Gs_1\Gs_2^2)^2=(\Gs_1\Gs_2)^3$ in $\BG3$.)

1900: 

1901: 

1902: \subsection{Proof of Theorem~\ref{th.proper}}\label{proof.proper}

1903: The fact that the perturbation epimorphisms

1904: $G_2',G_2''\onto\BG3/(\Gs_1\Gs_2)^3$ are proper is proved in Eyral,

1905: Oka~\cite{EyralOka}, where it is shown that the Alexander module

1906: of a sextic with the set of singularities

1907: $(2\bE_6\splus\bA_5)\splus\bA_2$ has a

1908: torsion summand $\Z_2\times\Z_2$, whereas the Alexander modules

1909: of all other groups

1910: listed in Theorem~\ref{th.group}

1911: %calculated in the paper

1912: can easily be shown to be $\Z[t]/(t^2-t+1)$. (In other words, the

1913: abelianization of the commutant of~$G_2'$ or $G_2''$ is equal to

1914: $\Z_2\times\Z_2\times\Z\times\Z$, and for all other groups it

1915: equals $\Z\times\Z$.)

1916: 

1917: The epimorphism

1918: $$

1919: \Gf_0\:G_0=\BG4/\Gs_2\Gs_1^2\Gs_2\Gs_3^2\onto\BG3/(\Gs_1\Gs_2)^3

1920: $$

1921: is considered in Oka, Pho~\cite{OkaPho}. One can observe that both

1922: braids $\Gs_2\Gs_1^2\Gs_2\Gs_3^2$ and $(\Gs_1\Gs_2)^3$ in the

1923: definition of the groups are pure, \ie, belong to the kernels of

1924: the respective canonical epimorphism

1925: $\BG{n}\onto\BG{n}/\Gs_1^2=\SG{n}$. Furthermore, $\Gf_0$ takes each

1926: of the standard generators $\Gs_1$, $\Gs_2$, $\Gs_3$ of~$\BG4$ to

1927: a conjugate of $\Gs_1$. Hence, the induced epimorphism

1928: $G_0/\Gf^{-1}(\Gs_1^2)=\SG4\onto\BG3/\Gs_1^2=\SG3$ is proper, and

1929: so is~$\Gf_0$.

1930: 

1931: A similar argument applies to the epimorphism

1932: $\Gf_3\:G_3\onto G_0$, which takes each generator $\Ga$, $\Gb$,

1933: $\bGb$ of~$G_3$ to a conjugate of~$\Gs_1\in G_0$. The induced

1934: epimorphism

1935: $$

1936: G_3/\Gf_3^{-1}(\Gs_1^2)=\SL(2,\Bbb F_3)\onto

1937:  G_0/\Gs_1^2=\SG4=\PSL(2,\Bbb F_3)

1938: $$

1939: is proper; hence, so is $\Gf_3$. (Alternatively, one can compare

1940: $G_3/\Gf_3^{-1}(\Gs_1^4)$ and $G_0/\Gs_1^4$, which are finite

1941: groups of order $3\cdot2^9$ and $3\cdot2^6$, respectively. The finite

1942: quotients of~$G_3$ and~$G_0$ were computed using {\tt GAP}~\cite{GAP}.)

1943: \qed

1944: 

1945: 

1946: \section{Perturbations\label{S.perturbations}}

1947: 

1948: \subsection{Perturbing a singular point}\label{s.perturbations}

1949: Consider a singular point~$P$ of a plain curve $B$ and a Milnor

1950: ball~$\MB$ around~$P$. Let~$B'$ be a nontrivial (\ie, not

1951: equisingular) perturbation of~$B$ such that,

1952: during the perturbation, the curve remains transversal to

1953: $\partial\MB$.

1954: 

1955: \lemma\label{pert.E6}

1956: In the notation above, let $P$ be of type~$\bE_6$. Then $B'\cap\MB$

1957: has one of the following sets of singularities\rom:

1958: \roster

1959: \item\local{E6.max}

1960: $2\bA_2\splus\bA_1$\rom: one has $\pi_1(\MB\sminus B')=\BG4$\rom;

1961: \item\local{E6.braid}

1962: $\bA_5$ or $2\bA_2$\rom: one has $\pi_1(\MB\sminus B')=\BG3$\rom;

1963: \item\local{E6.cyclic}

1964: $\bD_5$, $\bD_4$, $\bA_4\splus\bA_1$, $\bA_4$, $\bA_3\splus\bA_1$,

1965: $\bA_3$, $\bA_2\splus k\bA_1$ \rom($k=0$, $1$, or~$2$\rom), or

1966: $k\bA_1$ \rom($k=0$, $1$, $2$, or~$3$\rom)\rom:

1967: one has $\pi_1(\MB\sminus B')=\Z$.

1968: \endroster

1969: \endlemma

1970: 

1971: \proof

1972: The perturbations of a simple singularity are enumerated by the

1973: subgraphs of its Dynkin graph, see E.~Brieskorn~\cite{Brieskorn}

1974: or G.~Tjurina~\cite{Tjurina}.

1975: For the fundamental group, observe that

1976: the space $\MB\sminus B$ is diffeomorphic to

1977: $\Cp2\sminus(C\cup L)$, where $C\subset\Cp2$ is a plane quartic

1978: with a type~$\bE_6$ singular point, and $L$ is a line with a

1979: single quadruple intersection point with~$C$. Then, the

1980: perturbations of~$B$ inside~$\MB$ can be regarded as perturbations

1981: of~$C$ keeping the point of quadruple intersection with~$L$,

1982: see~\cite{quintics}, and the perturbed fundamental group

1983: $\pi_1(\Cp2\sminus(C'\cup L)\cong\pi_1(\MB\sminus B')$

1984: is found in~\cite{groups}.

1985: \endproof

1986: 

1987: \lemma\label{pert.A5}

1988: In the notation above, let $P$ be of type~$\bA_5$. Then $B'\cap\MB$

1989: has one of the following sets of singularities\rom:

1990: \roster

1991: \item\local{A5.braid}

1992: $2\bA_2$\rom: one has $\pi_1(\MB\sminus B')=\BG3$\rom;

1993: \item\local{A5.abelian}

1994: $\bA_3\splus\bA_1$ or $3\bA_1$\rom:

1995: one has $\pi_1(\MB\sminus B')=\Z\times\Z$\rom;

1996: \item\local{A5.cyclic}

1997: $\bA_4$, $\bA_3$, $\bA_2\splus\bA_1$, $\bA_2$, or

1998: $k\bA_1$ \rom($k=0$, $1$, or~$2$\rom)\rom:

1999: one has $\pi_1(\MB\sminus B')=\Z$.

2000: \endroster

2001: \endlemma

2002: 

2003: \lemma\label{pert.A2}

2004: In the notation above, let $P$ be of type~$\bA_2$. Then $B'\cap\MB$

2005: has the set of singularities $\bA_1$ or $\varnothing$, and

2006: one has $\pi_1(\MB\sminus B')=\Z$.

2007: \endlemma

2008: 

2009: \proof[Proof of Lemmas~\ref{pert.A5} and~\ref{pert.A2}]

2010: Both statements are a well known property of type~$\bA$ singular

2011: points: any perturbation of a type~$\bA_p$ singular point has the

2012: set of singularities $\bigsplus\bA_{p_i}$ with

2013: $d=(p+1)-\sum(p_i+1)\ge0$, and the group $\pi_1(\MB\sminus B')$ is

2014: given by $\<\Ga,\Gb\,|\,\Gs^s\Ga=\Ga,\ \Gs^s\Gb=\Gb>$, where $\Gs$

2015: is the standard generator of the braid group~$\BG2$ acting on

2016: $\<\Ga,\Gb>$ and $s=1$ if $d>0$ or $s=\gcd(p_i+1)$ if $d=0$.

2017: \endproof

2018: 

2019: \proposition\label{E6.onto}

2020: Let~$B$ be a plane sextic of torus type with at least two

2021: type~$\bE_6$ singularities, and let $\MB$ be a Milnor ball about a

2022: type~$\bE_6$ singular point of~$B$. Then the inclusion

2023: homomorphism $\pi_1(\MB\sminus B)\to\pi_1(\Cp2\sminus B)$

2024: is onto.

2025: \endproposition

2026: 

2027: \proof

2028: The proposition is an immediate consequence of

2029: Corollaries~\ref{e6->>pi}, \ref{e6->>pi.a3},

2030: \ref{e6->>pi.a5.1} and~\ref{e6->>pi.a5.2}.

2031: \endproof

2032: 

2033: \corollary\label{E6.perturbed}

2034: Let~$B$ be a plane sextic of torus type with at least two

2035: type~$\bE_6$ singular points, and let $B'$ be a perturbation of~$B$.

2036: \roster

2037: \item\local{E6.perturbed.abelian}

2038: If at least one of the type~$\bE_6$ singular points of~$B$ is

2039: perturbed as in~\iref{pert.E6}{E6.cyclic}, then

2040: $\pi_1(\Cp2\sminus B')=\CG6$.

2041: \item\local{E6.perturbed.braid}

2042: If at least one of the type~$\bE_6$ singular points of~$B$ is

2043: perturbed as in~\iref{pert.E6}{E6.braid}

2044: and $B'$ is still of torus type,

2045: then $\pi_1(\Cp2\sminus B')=\BG3/(\Gs_1\Gs_2)^3$.

2046: \endroster

2047: \endcorollary

2048: 

2049: \proof

2050: Let~$\MB$ be a Milnor ball about the type~$\bE_6$ singular point in

2051: question. Due to

2052: Proposition~\ref{E6.onto},

2053: the inclusion homomorphism

2054: $\pi_1(\MB\sminus B)\to\pi_1(\Cp2\sminus B)$ is onto. Hence, in

2055: case~\loccit{E6.perturbed.abelian}, there

2056: is an epimorphism $\Z\onto\pi_1(\Cp2\sminus B')$, and

2057: in case~\loccit{E6.perturbed.braid}, there

2058: is an epimorphism $\BG3\onto\pi_1(\Cp2\sminus B')$.

2059: In the former case, the epimorphism above implies that the group

2060: is abelian, hence~$\CG6$.

2061: In the latter case, the central element $(\Gs_1\Gs_2)^3\in\BG3$

2062: projects to $6\in\Z=\BG3/[\BG3,\BG3]$; since the abelianization of

2063: $\pi_1(\Cp2\sminus B')$ is $\CG6$, the epimorphism above must

2064: factor through an epimorphism

2065: $G:=\BG3/(\Gs_1\Gs_2)^3\onto\pi_1(\Cp2\sminus B')$. On the other

2066: hand, since $B'$ is assumed to be of torus type, there is an

2067: epimorphism $\pi_1(\Cp2\sminus B')\onto G$, and as

2068: $G\cong\PSL(2,\ZZ)$ is Hopfian (as it is obviously residually finite),

2069: each of the two epimorphisms is bijective.

2070: \endproof

2071: 

2072: \corollary\label{A5.perturbed}

2073: Let~$B$ be a plane sextic as in~\ref{s.2e6+a5+a2.1},

2074: and let $B'$ be a perturbation of~$B$ such that

2075: the type~$\bA_5$ singular point is

2076: perturbed as in~\iref{pert.A5}{A5.abelian} or~\ditto{A5.cyclic}.

2077: Then one has $\pi_1(\Cp2\sminus B')=\CG6$.

2078: \endcorollary

2079: 

2080: \proof

2081: %The statement follows from Proposition~\ref{a5->>pi} and

2082: %Lemma~\ref{pert.A5}.

2083: Due to Proposition~\ref{a5->>pi} and

2084: Lemma~\ref{pert.A5}, the group of the perturbed sextic~$B'$ is

2085: abelian. Since $B'$ is irreducible, $\pi_1(\Cp2\sminus B')=\CG6$.

2086: \endproof

2087: 

2088: \corollary\label{A2.perturbed}

2089: Let~$B$ be a plane sextic as in~\ref{s.2e6+2a2+a3},

2090: and let $B'$ be a perturbation of~$B$ such that

2091: an inner type~$\bA_2$ singular point of~$B$ is

2092: perturbed to~$\bA_1$ or~$\varnothing$.

2093: Then one has $\pi_1(\Cp2\sminus B')=\CG6$.

2094: \endcorollary

2095: 

2096: \proof

2097: Let~$P$ be the inner type~$\bA_2$ singular point perturbed, and

2098: let $\MB$ be a Milnor ball about~$P$. In the notation of

2099: Section~\ref{s.2e6+2a2+a3}, the group $\pi_1(\MB\sminus B)$ is

2100: generated by~$\Ga$ and~$\Gb$ (or $\bGa=\Ga$ and~$\bGb$ for the

2101: other point), and the perturbation results in an extra relation

2102: $\Ga=\Gb$. Then \eqref{eq.2.3} implies $\bGb=\Gb$ and the group is

2103: cyclic.

2104: \endproof

2105: 

2106: %$\<-1>\oplus\<1>\oplus\<1>$: $e_1=-v_1+v_2$, $e_2=-v_2$,

2107: %$e_3=v_0+v_1+v_2$.

2108: %

2109: %$t_{e_1}\:v_1\leftrightarrow v_2$;\quad

2110: %$t_{e_2}\:v_2\mapsto-v_2$;\quad

2111: %$t_{e_3}\:v_0\mapsto v_1+v_2$, $v_1\mapsto v_0-v_1+v_2$,

2112: %$v_2\mapsto v_0+v_1-v_2$. Let $u_0=v_0$, $u_1=v_1+v_2$,

2113: %$u_2=v_1-v_2$. Then $t_{e_2}\:u_1\leftrightarrow u_2$ and

2114: %$t_{e_3}\:u_0\mapsto u_1$, $u_1\mapsto-u_0$,

2115: %$u_2\mapsto u_2$.

2116: %

2117: %$\<-2>\oplus\<1>\oplus\<1>$: $e_1=-v_1+v_2$, $e_2=-v_2$,

2118: %$e_3=v_0+2v_1$.

2119: %

2120: %$t_{e_1}\:v_1\leftrightarrow v_2$;\quad

2121: %$t_{e_2}\:v_2\mapsto-v_2$;\quad

2122: %$t_{e_3}\:v_0\mapsto3v_0+4v_1\equiv v_1$,

2123: %$v_1\mapsto-2v_0-3v_1\equiv v_0$,

2124: %$v_2\mapsto v_2$.

2125: 

2126: \midinsert

2127: \table\label{tab.nontorus}

2128: Sextics with abelian fundamental group

2129: \endtable

2130: \hbox to\hsize{\hss

2131: \TAB\NonTorus

2132: 2 e[6] + d[5] + a[1],               [37, 4], [2, 2, 0], 18, [3, 2, false]

2133: 2 e[6] + a[4] + 2 a[1],             [38, 4], [2, 2, 1], 18, [3, 2, false]

2134: 2 e[6] + a[3] + a[2] + a[1],        [40, 5], [2, 3, 0], 18, [3, 3, false]

2135: 2 e[6] + 2 a[2] + a[1],             [38, 6], [1, 4, 0], 17, [4, 4, false]

2136: 2 e[6] + a[2] + 3 a[1],             [36, 5], [3, 3, 0], 17, [4, 3, false]

2137: e[6] + 2 d[5] + a[1],               [36, 2], [3, 1, 0], 17, [4, 3, false]

2138: e[6] + d[5] + a[5] + a[1],          [37, 4], [3, 2, 0], 17, [4, 3, false]

2139: e[6] + d[5] + a[4] + 2 a[1],        [37, 2], [3, 1, 1], 17, [4, 3, false]

2140: e[6] + d[5] + a[3] + 2 a[2],        [43, 4], [2, 3, 0], 18, [3, 3, false]

2141: e[6] + d[5] + 3 a[2],               [41, 5], [1, 4, 0], 17, [4, 4, false]

2142: e[6] + d[5] + 2 a[2] + 2 a[1],      [39, 4], [3, 3, 0], 17, [4, 3, false]

2143: e[6] + d[4] + a[3] + 2 a[2],        [41, 4], [3, 3, 0], 17, [4, 3, false]

2144: e[6] + d[4] + 3 a[2],               [39, 5], [2, 4, 0], 16, [5, 4, false]

2145: e[6] + d[4] + 2 a[2] + 2 a[1],      [37, 4], [4, 3, 0], 16, [5, 4, false]

2146: e[6] + a[5] + a[4] + a[1] + a[2],   [42, 5], [2, 3, 1], 18, [3, 3, false]

2147: e[6] + a[5] + a[3] + a[2] + a[1],   [40, 5], [3, 3, 0], 17, [4, 3, false]

2148: e[6] + 2 a[4] + 3 a[1],             [38, 2], [3, 1, 2], 17, [4, 3, false]

2149: e[6] + a[4] + a[3] + 2 a[2] + a[1], [44, 4], [2, 3, 1], 18, [3, 3, false]

2150: e[6] + a[4] + 2 a[2] + 3 a[1],      [40, 4], [3, 3, 1], 17, [4, 3, false]

2151: e[6] + 2 a[3] + 2 a[2] + a[1],      [42, 4], [3, 3, 0], 17, [4, 3, false]

2152: 2 d[5] + a[5] + a[1],               [36, 2], [4, 1, 0], 16, [5, 4, false]

2153: 2 d[5] + a[4] + 2 a[1],             [36, 0], [4, 0, 1], 16, [5, 4, false]

2154: 2 d[5] + a[3] + 2 a[2],             [42, 2], [3, 2, 0], 17, [4, 3, false]

2155: 2 d[5] + 3 a[2],                    [40, 3], [2, 3, 0], 16, [5, 3, true]

2156: 2 d[5] + 2 a[2] + 2 a[1],           [38, 2], [4, 2, 0], 16, [5, 4, false]

2157: d[5] + d[4] + a[3] + 2 a[2],        [40, 2], [4, 2, 0], 16, [5, 4, false]\ENDTAB

2158: \hss

2159: \TAB\NonTorus

2160: d[5] + d[4] + 3 a[2],               [38, 3], [3, 3, 0], 15, [6, 3, true]

2161: d[5] + d[4] + 2 a[2] + 2 a[1],      [36, 2], [5, 2, 0], 15, [6, 5, false]

2162: d[5] + 2 a[5] + a[1],               [37, 4], [4, 2, 0], 16, [5, 4, false]

2163: d[5] + a[5] + a[4] + 2 a[1],        [37, 2], [4, 1, 1], 16, [5, 4, false]

2164: d[5] + a[5] + a[3] + 2 a[2],        [43, 4], [3, 3, 0], 17, [4, 3, false]

2165: d[5] + 2 a[4] + 3 a[1],             [37, 0], [4, 0, 2], 16, [5, 4, false]

2166: d[5] + a[4] + a[3] + 2 a[2] + a[1], [43, 2], [3, 2, 1], 17, [4, 3, false]

2167: d[5] + a[4] + 2 a[2] + 3 a[1],      [39, 2], [4, 2, 1], 16, [5, 4, false]

2168: d[5] + 2 a[3] + 2 a[2] + a[1],      [41, 2], [4, 2, 0], 16, [5, 4, false]

2169: 2 d[4] + a[3] + 2 a[2],             [38, 2], [5, 2, 0], 15, [6, 5, false]

2170: 2 d[4] + 3 a[2],                    [36, 3], [4, 3, 0], 14, [7, 4, true]

2171: d[4] + a[5] + a[3] + 2 a[2],        [41, 4], [4, 3, 0], 16, [5, 4, false]

2172: d[4] + a[4] + a[3] + 2 a[2] + a[1], [41, 2], [4, 2, 1], 16, [5, 4, false]

2173: d[4] + a[4] + 2 a[2] + 3 a[1],      [37, 2], [5, 2, 1], 15, [6, 5, false]

2174: d[4] + 2 a[3] + 2 a[2] + a[1],      [39, 2], [5, 2, 0], 15, [6, 5, false]

2175: 2 a[5] + a[4] + a[1] + a[2],        [42, 5], [3, 3, 1], 17, [4, 3, false]

2176: 2 a[5] + a[3] + a[2] + a[1],        [40, 5], [4, 3, 0], 16, [5, 4, false]

2177: a[5] + 2 a[4] + 2 a[1] + a[2],      [42, 3], [3, 2, 2], 17, [4, 3, false]

2178: a[5] + a[4] + a[3] + 2 a[2] + a[1], [44, 4], [3, 3, 1], 17, [4, 3, false]

2179: a[5] + 2 a[3] + 2 a[2] + a[1],      [42, 4], [4, 3, 0], 16, [5, 4, false]

2180: 3 a[4] + 4 a[1],                    [38, 0], [4, 0, 3], 16, [5, 4, false]

2181: 2 a[4] + a[3] + 2 a[2] + 2 a[1],    [44, 2], [3, 2, 2], 17, [4, 3, false]

2182: 2 a[4] + 2 a[2] + 4 a[1],           [40, 2], [4, 2, 2], 16, [5, 4, false]

2183: a[4] + 2 a[3] + 2 a[2] + 2 a[1],    [42, 2], [4, 2, 1], 16, [5, 4, false]

2184: 3 a[3]+2 a[2]+2 a[1],               [40, 2], [5, 2, 0], 15, [6, 5, false]\ENDTAB\hss}

2185: %% 244 types (not covered by Nori)

2186: %% 117 types not covered by 8a2.tex

2187: \endinsert

2188: 

2189: \subsection{Abelian perturbations}

2190: Theorem~\ref{th.nontorus} below lists the sets of singularities

2191: obtained by perturbing at least one inner singular point from a

2192: set listed in Table~\ref{tab.list}, not covered by Nori's

2193: theorem~\cite{Nori}, and not appearing in~\cite{degt.8a2}.

2194: 

2195: \theorem\label{th.nontorus}

2196: Let $\Sigma$ be a set of singularities obtained from one of those

2197: listed in Table~\ref{tab.nontorus} by several \rom(possibly

2198: none\rom) perturbations $\bA_2\to\bA_1,\varnothing$ or

2199: $\bA_1\to\varnothing$.

2200: Then $\Sigma$ is realized by an irreducible plane

2201: sextic, not of torus type, whose fundamental group is~$\CG6$.

2202: \endtheorem

2203: 

2204: Altogether, perturbations as in Theorem~\ref{th.nontorus} produce

2205: $244$ sets of singularities not covered by Nori's theorem; $117$

2206: of them are new as compared to~\cite{degt.8a2}.

2207: 

2208: \proof

2209: Each set of singularities in question is obtained by a perturbation

2210: from one of the sets of singularities listed in

2211: Table~\ref{tab.list}. Furthermore, the perturbation can be chosen

2212: so that at least one type~$\bE_6$ singular point is perturbed as

2213: in~\iref{pert.E6}{E6.cyclic}, or the type~$\bA_5$ singular point

2214: is perturbed as in~\iref{pert.A5}{A5.cyclic}, or at least one

2215: inner cusp is perturbed to~$\bA_1$ or~$\varnothing$. According

2216: to~\cite{degt.8a2}, any such (formal) perturbation is realized by

2217: a family of sextics, and due to

2218: Corollaries~\iref{E6.perturbed}{E6.perturbed.abelian},

2219: \ref{A5.perturbed}, and~\ref{A2.perturbed}, the perturbed sextic

2220: has abelian fundamental group.

2221: \endproof

2222: 

2223: 

2224: \subsection{Non-abelian perturbations}

2225: In this section, we treat the few perturbations of torus type that

2226: can be obtained from Table~\ref{tab.list} and do not appear

2227: in~\cite{degt.8a2}.

2228: 

2229: \theorem\label{th.torus}

2230: Each of the eight sets of singularities listed in

2231: Table~\ref{tab.torus} is realized by an irreducible plane sextic

2232: of torus type whose fundamental group is $\BG3/(\Gs_1\Gs_2)^3$.

2233: \midinsert

2234: \table\label{tab.torus}

2235: Sextics of torus type

2236: \endtable

2237: \hbox to\hsize{\hss

2238: \Tab

2239: (\bE_6\splus2\bA_5)\splus\bA_2\cr

2240: (\bE_6\splus2\bA_5)\splus\bA_1\cr

2241: (\bE_6\splus2\bA_5)\cr

2242: (3\bA_5)\splus\bA_2

2243: \ENDTAB\hss\Tab

2244: (3\bA_5)\splus\bA_1\cr

2245: (3\bA_5)\cr

2246: (\bE_6\splus\bA_5\splus2\bA_2)\splus\bA_3\cr

2247: (2\bA_5\splus2\bA_2)\splus\bA_3

2248: \ENDTAB

2249: \hss}

2250: %% 8 sets

2251: \endinsert

2252: \endtheorem

2253: 

2254: Theorem~\ref{th.torus} covers two tame sextics:

2255: $(\bE_6\splus2\bA_5)$ and $(3\bA_5)$. The fundamental groups of these

2256: curves were first found in Oka, Pho~\cite{OkaPho}.

2257: 

2258: \proof

2259: As in the previous section, we perturb one of the sets of

2260: singularities listed in Table~\ref{tab.list}, this time making

2261: sure that

2262: \roster

2263: \item\local{pert.1}

2264: each type~$\bE_6$ singular point is perturbed as

2265: in~\iref{pert.E6}{E6.max} or~\ditto{E6.braid} (or is not

2266: perturbed at all),

2267: \item\local{pert.2}

2268: each type~$\bA_5$ singular point is perturbed as

2269: in~\iref{pert.A5}{A5.braid} (or is not

2270: perturbed at all),

2271: \item\local{pert.3}

2272: none of the inner cusps is perturbed, and

2273: \item\local{pert.4}

2274: at least one type~$\bE_6$ singular point is perturbed as

2275: in~\iref{pert.E6}{E6.braid}.

2276: \endroster

2277: (Note that, in the case under consideration, inner are the cusps

2278: appearing from the cusp of~$\B$.)

2279: From the arithmetic description of curves of torus type given

2280: in~\cite{degt.Oka} (see also~\cite{JAG}) it follows that any

2281: perturbation satisfying~\loccit{pert.1}--\loccit{pert.3} above

2282: preserves the torus structure; then, in view of~\loccit{pert.4},

2283: Corollary~\iref{E6.perturbed}{E6.perturbed.braid} implies that the

2284: resulting fundamental group is $\BG3/(\Gs_1\Gs_2)^3$.

2285: \endproof

2286: 

2287: \Remark\label{not.new}

2288: If all type~$\bE_6$ singular points are perturbed as

2289: in~\iref{pert.E6}{E6.max} (or not perturbed at all), the study of

2290: the fundamental group would require more work; in particular, one

2291: would need an explicit description of the homomorphism

2292: $\pi_1(\MB_{\bE_6}\sminus B)\onto\pi_1(\MB_{\bE_6}\sminus B')$. On

2293: the other hand, it is easy to show that such perturbations do not

2294: give anything new compared to~\cite{degt.8a2}. (In fact,

2295: using~\cite{JAG}, one can even show that the deformation classes

2296: of the sextics obtained are the same; it suffices to prove the

2297: connectedness of the deformation families realizing the

2298: sets of singularities

2299: $(\bE_6\splus\bA_5\splus2\bA_2)\splus\bA_2\splus\bA_1$ and

2300: $(\bE_6\splus4\bA_2)\splus\bA_3\splus\bA_1$,

2301: which are maximal in the context.) For this reason, we do not

2302: consider these perturbations here.

2303: \endRemark

2304: 

2305: 

2306: \widestnumber\key{EO1}

2307: \refstyle{C}

2308: \widestnumber\no{99}

2309: \Refs

2310: 

2311: \ref{Bri}

2312: \by E.~Brieskorn

2313: \paper Singular elements of semi-simple algebraic groups

2314: \inbook Actes Congr\`es Inter. Math., Nice

2315: \issue 2

2316: \yr 1970

2317: \pages 279-284

2318: \endref\label{Brieskorn}

2319: 

2320: \ref{D1}

2321: \by A.~Degtyarev

2322: \paper Isotopy classification of complex plane projective curves of

2323: degree~$5$

2324: \jour Algebra i Analis

2325: \yr 1989

2326: \vol    1

2327: \issue  4

2328: \pages  78--101

2329: \lang Russian

2330: \moreref\nofrills English transl. in

2331: \jour Leningrad Math.~J.

2332: \vol 1

2333: \yr 1990

2334: \issue 4

2335: \pages 881--904

2336: \endref\label{quintics}

2337: 

2338: \ref{D2}

2339: \by A.~Degtyarev

2340: \paper Quintics in $\C\roman{p}^2$ with nonabelian fundamental group

2341: \jour Algebra i Analis

2342: \yr 1999

2343: \vol    11

2344: \issue  5

2345: \pages  130--151

2346: \lang Russian

2347: \moreref\nofrills English transl. in

2348: \jour Leningrad Math.~J.

2349: \vol 11

2350: \yr 2000

2351: \issue 5

2352: \pages 809--826

2353: \endref\label{groups}

2354: 

2355: \ref{D3}

2356: \by A.~Degtyarev

2357: \paper On deformations of singular plane sextics

2358: \jour J. Algebraic Geom.

2359: \vol 17

2360: \yr 2008

2361: \pages 101--135

2362: \endref\label{JAG}

2363: 

2364: \ref{D4}

2365: \by A.~Degtyarev

2366: \paper Oka's conjecture on irreducible plane sextics

2367: \jour J. London Math. Soc.

2368: \toappear

2369: \finalinfo\tt arXiv:\allowbreak math.AG/0701671

2370: \endref\label{degt.Oka}

2371: 

2372: \ref{D5}

2373: \by A.~Degtyarev

2374: \paper Zariski $k$-plets via dessins d'enfants

2375: \jour Comment. Math. Helv.

2376: \finalinfo\tt arXiv:0710.0279

2377: \toappear

2378: \endref\label{degt.kplets}

2379: 

2380: \ref{D5}

2381: \by A.~Degtyarev

2382: \paper On irreducible sextics with non-abelian fundamental group

2383: \inbook Fourth Franco-Japanese Symposium on Singularities (Toyama, 2007)

2384: \finalinfo\tt arXiv:0711.3070

2385: \toappear

2386: \endref\label{degt.Oka3}

2387: 

2388: \ref{D6}

2389: \by A.~Degtyarev

2390: \paper Irreducible plane sextics with large fundamental groups

2391: \finalinfo\tt arXiv:0712.2290

2392: \endref\label{degt.8a2}

2393: 

2394: \ref{D7}

2395: \by A.~Degtyarev

2396: \paper Stable symmetries of plane sextics

2397: \finalinfo\tt arXiv:0802.2336

2398: \endref\label{symmetric}

2399: 

2400: \ref{EO}

2401: \by C.~Eyral, M.~Oka

2402: \paper Fundamental groups of the complements of certain plane

2403: non-tame torus sextics

2404: \jour Topology Appl.

2405: \vol 153

2406: \yr 2006

2407: \issue 11

2408: \pages 1705--1721

2409: \endref\label{EyralOka}

2410: 

2411: \ref{GAP}

2412: \by The GAP Group

2413: \book GAP --- Groups, Algorithms, and Programming

2414: \bookinfo Version 4.4.10

2415: \yr 2007

2416: \finalinfo ({\tt http://www.gap-system.org})

2417: \endref\label{GAP}

2418: 

2419: \ref{vK}

2420: \by E.~R.~van~Kampen

2421: \paper On the fundamental group of an algebraic curve

2422: \jour  Amer. J. Math.

2423: \vol   55

2424: \yr    1933

2425: \pages 255--260

2426: \endref\label{vanKampen}

2427: 

2428: \ref{No}

2429: \by M.~V.~Nori

2430: \paper Zariski conjecture and related problems

2431: \jour Ann. Sci. \'Ec. Norm. Sup., 4 s\'erie

2432: \vol    16

2433: \yr     1983

2434: \pages  305--344

2435: \endref\label{Nori}

2436: 

2437: \ref{OP1}

2438: \by M.~Oka, D.~T.~Pho

2439: \paper Classification of sextics of torus type

2440: \jour Tokyo J. Math.

2441: \vol 25

2442: \issue 2

2443: \pages 399--433

2444: \yr 2002

2445: \endref\label{OkaPho.moduli}

2446: 

2447: \ref{OP2}

2448: \by M.~Oka, D.~T.~Pho

2449: \paper Fundamental group of sextics of torus type

2450: \inbook Trends in singularities

2451: \pages 151--180

2452: \bookinfo Trends Math.

2453: \publ Birkh\"auser

2454: \publaddr Basel

2455: \yr 2002

2456: \endref\label{OkaPho}

2457: 

2458: \ref{Oz}

2459: \by A.~\"Ozg\"uner

2460: \book Classical Zariski pairs with nodes

2461: \bookinfo M.Sc. thesis

2462: \publ Bilkent University

2463: \yr 2007

2464: \endref\label{Aysegul}

2465: 

2466: \ref{Tju}

2467: \by G.~Tjurina

2468: \paper Resolution of singularities of flat deformations of double rational

2469: points

2470: \jour Funktsional Anal. i Pril.

2471: \vol 4

2472: \yr 1970

2473: \issue 1

2474: \pages 77--83

2475: \lang Russian

2476: \transl\nofrills English transl. in

2477: \jour Functional Anal. Appl.

2478: \vol 4

2479: \issue 1

2480: \yr 1970

2481: \pages 68-73

2482: \endref\label{Tjurina}

2483: 

2484: \ref{T}

2485: \by H.~Tokunaga

2486: \paper $(2,3)$-torus sextics and the Albanese images of $6$-fold

2487: cyclic multiple planes

2488: \jour Kodai Math.~J.

2489: \vol 22

2490: \yr 1999

2491: \issue 2

2492: \pages 222--242

2493: \endref\label{Tokunaga}

2494: 

2495: \ref{Z}

2496: \by O.~Zariski

2497: \paper On the problem of existence of algebraic functions of two

2498: variables possessing a given branch curve

2499: \jour Amer. J. Math.

2500: \vol 51

2501: \yr 1929

2502: \pages 305--328

2503: \endref\label{Zariski}

2504: 

2505: \ref{Ya}

2506: \by J.-G.~Yang

2507: \paper Sextic curves with simple singularities

2508: \jour Tohoku Math. J. (2)

2509: \vol 48

2510: \issue 2

2511: \yr 1996

2512: \pages 203--227

2513: \endref\label{Yang}

2514: 

2515: \endRefs

2516: 

2517: \enddocument

2518: