1: \documentclass[12pt,leqno]{article}

2: \oddsidemargin  = 9mm

3: \evensidemargin = 0 cm

4: \textwidth      = 14.8 cm

5: \textheight     = 22 cm

6: \headheight     = 0 cm

7: \topskip        = 0 cm

8: \topmargin      = 0 cm

9: %\pagestyle{myheadings}

10: \usepackage{graphicx}

11: \input amssym.tex

12: %\input alinea.tex

13: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%WRITTEN IN LATEX

14: \begin{document}

15: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%DEFINITIONS

16: \def\diamond{\square}

17: \def\pn{{\Bbb C}P(N)}

18: \def\pnh{\check{{\Bbb C}P(N)}}

19: \def\Sf{{\cal F}(2,d)}   	 

20: \def\C{\Bbb C}

21: \def\F{{\cal F}}	  		

22: \def\pr{{\bf Proof: }}

23: \def\S{\overline{\C}}

24: \def\P{{\cal P}}

25: \def\RR{\{F=0\}\cap\{G=0\}}

26: \def\INT{pGdF-qFdG}

27: \def\integ{\frac{F^p}{G^q}}

28: \def\integb{\frac{\bar{F}^p}{\bar{G}^q}}

29: \def\PP{{\cal P}_a \times {\cal P}_b}

30: \def\t{\tilde}

31: \def\In{{\cal I}(n,a,b)}

32: \def\Inm{{\cal I} _m(n,a,b)} 

33: \def\pro{{\Bbb C}P(n)}

34: \def\tpro{\tilde{{\Bbb C}P(n)}}

35: \def\Z{{\Bbb Z}}

36: \def\N{{\Bbb N}}

37: \def\H{{\cal H}}

38: \def\R{{\cal R}}

39: \def\integr{\frac{\t{F}^\frac{1}{q}}{\t{G}^\frac{1}{p}}}

40: 

41: \newtheorem{theo}{Theorem}[subsection]

42: \newtheorem{Theo}{Theorem}

43: \newtheorem{defi}{Definition}[subsection]

44: \newtheorem{coro}{Corollary}[subsection]

45: \newtheorem{lem}{Lemma}[subsection]

46: \newtheorem{claa}{Proposition}[section]

47: \newtheorem{exa}{Example}[subsection]

48: \newtheorem{rem}{Remark}[section]

49: \begin{center}

50: {\LARGE\bf On the Topology of Foliations with a First Integral\\}

51: \vspace{.25in}

52: {\large {\sc Hossein Movasati}\footnote{Partially supported by CNPq-Brazil

53: \\

54: Keywords: Lefschetz pencil and vanishing cycle - homology groups - fiber bundle

55: \\

56: Math. classification: 14D99 - 57R30}

57: %\\

58: %Instituto de Matem\'atica Pura e Aplicada, IMPA

59: %\\

60: %Estrada Dona Castorina, 110, 22460-320

61: %\\

62: %Rio de Janeiro, RJ, Brazil

63: %\\

64: %E-mail: hossein@impa.br\\

65: %July 2000}

66: }

67: %\setcounter{page}{33} 

68: \end{center}

69: \begin{abstract}

70: The main objective of this article is to study the topology of the fibers of a generic rational function of the type $\integ$ in the projective space of dimension two. We will

71: prove that the action of the monodromy group on a single Lefschetz vanishing cycle $\delta$ generates the first homology group of a generic fiber of $\integ$.

72: In particular, we will prove that for any two Lefschetz vanishing cycles $\delta_0$ and $\delta_1$ in 

73: a regular compact fiber of $\integ$, there exists a monodromy $h$ such that

74: $h(\delta_0)=\pm \delta_1$. 

75: \end{abstract}

76: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%ABSTRACT

77: \setcounter{section}{-1}

78: \section{Introduction}

79: \ \ \ \ \

80: Let $F$ and $G$ be two homogeneous polynomials in $\C^{n+1}$. The following

81: function  is well-defined

82: \[

83: f=\integ :\pro\backslash\R \rightarrow \S

84: \]

85: \[

86: f(x)=

87: \frac{F(x)^p}{G(x)^q},\ x=[x_0;x_1;\cdots ;x_n]

88: \]

89: where $\R=\{F=0\}\cap\{G=0\}$, $\frac{deg(F)}{deg(G)}=\frac{q}{p}$ and $p$ and

90: $q$ are relatively prime numbers. We can view the fibration of $f$ as a

91: codimension one foliation in $\pro$ given by the 1-form

92: \[

93: \omega=\INT

94: \]

95: Let $\P_a$ denote the set of homogeneous polynomials of degree $a$ in $\C^n$. 

96: \begin{claa}

97: \label{gen}

98: There exists an open dense subset $U$ of $\P_a\times\P_b$ such that for any $(F,G)\in U$ we have:

99: \begin{enumerate}

100: \item

101: $\{F=0\}$ and $\{G=0\}$ are smooth varieties in $\pro$ and intersect each other 

102: transversally;

103: \item

104: The restriction of $f$ to ${\pro\backslash (\{F=0\}\cup \{G=0\})}$ has 

105: nondegenerate critical points, namely $p_1,p_2,\ldots,p_r$, 

106: with distinct images in $\S$, namely

107: $c_1,c_2,\ldots,c_r$ respectively.

108: \end{enumerate}

109: \end{claa}

110: Throughout the text the elements of $U$ will be called generic elements. We will prove this proposition in Appendix ~\ref{secgeneric}. Put 

111: \[

112: C=\{c_1,c_2,\ldots,c_r\}

113: \]

114: From now on, we will work with the function $f$ which has the generic

115:  properties as in

116: Proposition ~\ref{gen}. The foliation $\F$ associated to $f$ has the following

117: singular set

118: \[

119: Sing(\F)=\{p_1,p_2,\ldots,p_r,\R\}

120: \]

121: The value $0$ (resp. $\infty$) is a critical value of $f$ if and only if $p>1$ (resp. $q>1$). Let $A$ be a subset of $\{0,\infty\}$ which consists of only critical values.

122: For example if $p=1,q=1$ then $A$ is empty. The set of critical values of $\integ$

123: is $C\cup A$.  

124: \begin{claa}

125: $f$ is a $C^\infty$ fiber bundle map over $\S\backslash (C\cup A)$.

126: \end{claa}

127: We will prove this proposition in Section ~\ref{top}.

128: \\

129: 

130: The above proposition enables us to use the arguments of Picard-Lefschetz Theory

131: to study the topology of the fibers of $f$. But, for example, the critical fiber

132: $\{F=0\}$, when $p>1$, is not considered in that theory (as far as I know). To

133: overcome with this obstacle, we will construct a ramification map 

134: $\tau:\tpro\rightarrow\pro$ for the multivalued function $f^{\frac{1}{pq}}$. The

135: pull-back function $\t{f}^{\frac{1}{pq}}=f^{\frac{1}{pq}}\circ\tau$ of 

136: $f^{\frac{1}{pq}}$ is univalued and has no more the critical values $0$

137: and $\infty$. Next, we will embed the complex manifold $\tpro$ in some $\pn$ in

138: such a way that the pull-back foliation $\F$ is obtained by the 

139: intersection of the

140: hyperplanes of a generic Lefschetz pencil with $\tpro$.

141: \\

142: The study of the topology of an algebraic variety by intersecting it with

143: hyperplanes of a pencil has been started systematicly by Lefschetz in his famous

144: article \cite{lef}. We will use the arguments of this area of mathematics, specially the articles \cite{lam},\cite{che}, to

145: understand the topology of the leaves of $\F$. Note that the leaves of $\F$ do

146: not contain the points of the set $\{F=0\}\cap\{G=0\}$.   

147: \\

148: In the first section we will construct such ramification map $\tau$, and in the

149: second section we will review Picard-Lefschetz Theory. In the third section

150: we will apply our results to the foliation $\F$.

151: \\

152:  I want to use this opportunity to express my thanks to my advisor

153:  Alcides Lins Neto. Discussions with him during this work have provided valuable information 

154:  and new ideas. Also, thanks go to Cesar Camacho and Paulo Sad for their

155:  interest and support. I also thank Eduardo Esteves for his comments on algebraic geometry.  The author is also grateful to the exceptional 

156:  scientific atmosphere in IMPA that made this work possible.

157: 

158:    

159: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%INTRODUCTION

160: \section{Ramification Maps}

161: \newtheorem{cla}{Proposition}[subsection]

162: \label{ramsec}

163: \ \ \ \ \

164: In the first part of this section we will introduce ramification maps with a

165:  normal crossing divisor. In the second part we will use a method which gives

166:   us some examples of ramification maps. This method will be enough for

167:   our purpose.

168: The following isomorphism will be used frequently during this section:

169: 

170: {\bf Leray (or Thom-Gysin) Isomorphism:} If a closed submanifold $N$ has 

171: pure real codimension $c$ in $M$, then there is an isomorphism 

172: \[

173: \tau :H_{k-c}(N)\tilde{\rightarrow}H_k(M,M\backslash N)

174: \]

175:  holding for any k, with the convention that $H_s(N)=0$ for

176:   $s<0$. Roughly speaking, given $x\in H_{k-c}(N)$, its image by this 

177:   isomorphism is obtained by thickening a cycle representing $x$, 

178:   each point of it growing into a closed $c$-disk transverse to $N$ in $M$ 

179:   (see \cite{che} p. 537). 

180:   \\

181:   Let $N$ be a connected codimension one submanifold of the complex manifold 

182:   $M$. Write the long exact sequence of the pair $(M,M\backslash N)$ as 

183:   follows:

184: \begin{equation}

185: \label{longler}

186: \cdots\rightarrow H_2(M,M\backslash N)

187: \stackrel{\sigma}{\rightarrow}H_1(M\backslash N)

188: \stackrel{i}{\rightarrow} H_1(M)\rightarrow \cdots

189: \end{equation}

190: where $\sigma$ is the boundary operator and $i$ is induced by inclusion. 

191: Since $N$ has real codimension two in $M$,  $H_2(M,M\backslash N)$ 

192: ($\simeq H_0(N)\simeq\Z$) is generated by the disk $\Delta$ transverse to 

193: $N$ at a point $y\in N$. By the above long exact sequence it follows that 

194: if a closed cycle $x$ in $M\backslash N$ is homologous to zero in $M$ then

195:  it is homologous to a multiple of $\sigma (\Delta)=\delta$ in $M\backslash D$. 

196: The cycle $\delta$ is called a simple loop around the point $y\in N$ in

197:  $M\backslash N$.  

198: \subsection{Normal Crossing Divisors} 

199: \ \ \ \ \ 

200:  The following well-known fact will be used frequently:

201: \begin{cla}

202: \label{cover}

203: Let $\tau :\t{M}\rightarrow M$ be a finite covering map of degree $p$. 

204: Then the following statements are true:

205: \begin{enumerate}

206: \item

207:  $\tau _* :\pi_1(\t{M})\rightarrow \pi_1(M)$ is one to one, where $\pi_1(M)$

208:  denotes the fundamental group of $M$;

209: \item

210:  If $\pi_1(M)$ is abelian then $\pi_1(\t{M})$ is also abelian and 

211:  $\pi_1(M)^p\subset \tau_*(\pi_1(\t{M}))$, where $\pi_1(M)^p=\{\gamma ^p\mid\gamma\in\pi_1(M) \}$.

212: \end{enumerate}

213: \end{cla}

214: \pr 

215: The proof of the first statement can be found in \cite{ste}. 

216: If $\pi_1(M)$ is abelian then  

217: \[

218: \tau_*(aba^{-1}b^{-1})=\tau_*(a)\tau_*(b)\tau_*(a)^{-1}\tau_*(b)^{-1}=1

219: \]

220: where $a,b\in\pi_1(\t{M})$. The map $\tau_*$ is one to one and so 

221: $aba^{-1}b^{-1}=1$ which implies that $ab=ba$.

222: \\

223: For any closed path $a\in\pi_1(M,x)$, its inverse image by $\tau$ is a 

224: union of closed paths $a_1,a_2,\ldots,a_k$ in $\t{M}$. 

225: Choose a point $y$ in $\t{M}$ and points $x_i$ in the path $a_i$, 

226: for $i=1,\ldots, k$ such that $\tau(x_i)=\tau(y)=x$ and put

227:  $b_i=A_i^{-1}a_iA_i$ and $b=b_1b_2\cdots b_k$, where $A_i$ is a 

228:  path in $\t{M}$ which connects $y$ to $x_i$. The image of $A_i$ by $\tau$ is a 

229:  closed path in $M$ and $\pi_1(M)$ is abelian, therefore

230: \[

231: \tau_*(b)=\Pi\tau_*(A_i)^{-1}\tau_*(a_i)\tau_*(A_i)=\Pi \tau_*{a_i}=a^p

232: \]

233: The last equality is true because $\tau$ has degree $p$ and the paths $a_i$'s are 

234: the inverse image of $a$.$\diamond$

235: \\

236: 

237: In what follows, if the considered group is abelian, we use the additive notations 

238: of groups; for example instead of $a^p$ we write $pa$.  

239:     

240: \begin{defi}\rm 

241: Let $M$ be a complex manifold of dimension $n$. 

242: By a reduced normal crossing divisor we mean a union of finitely many 

243: connected closed submanifolds, namely  $D_1,D_2,\ldots,D_s$ of $M$ and of 

244: codimension one, which intersect each other transversally i.e., 

245: for any point $a\in M$ there is a local coordinate 

246: $(x,y)\in \C^k\times \C^{n-k},\ k\leq s$ around $a$ such that in this coordinate 

247: $a=(0,0)$ and for any $j=1,\ldots ,k$, the component $D_{i_j}$ is given by 

248: $x_j=0$,

249:  where $\{i_1,\ldots,i_k\}=\{ i\mid a\in D_i\}$. We say this coordinate 

250:  normalizing coordinate of $D$ at $a$.

251:   We will denote a reduced normal crossing divisor by 

252:   \[

253:   D=\sum_1^s D_i

254:   \]

255: When $D$ has only one component i.e., $s=1$, then $D$ is called simple. 

256: \end{defi}

257: The following fact is a direct consequence of the definition.

258:  \begin{cla}

259: \label{subncd}

260: Let $D$ be a reduced normal crossing divisor in a complex manifold  $M$. 

261: For any  subset $I$ of $\{1,2,\ldots,s\}$, the set 

262: \[

263: M_I=\cap_{i\in I} D_i

264: \]

265: is a complex manifold of codimension $\# I$ in $M$ and 

266: \[

267: D_I=D\cap M_I=\sum_{i\not\in I} D_i\cap M_I

268: \]

269: is a reduced normal crossing divisor in $M_I$.   

270: \end{cla}

271: 

272: In what follows for a given function $\tau :\t{M}\rightarrow M$,

273: and for any subset $x$ of $M$ (resp. a meromorphic function $x$ on $M$) 

274: we denote by $\t{x}$ the set $\tau^{-1}(x)$ 

275: (resp. the meromorphic function $x\circ\tau$ on $\t{M}$).

276:   

277: \begin{defi}\rm

278: \label{ncd}

279: Let $M$ and $\t{M}$ be two compact complex manifolds of the same dimension,

280:  $D=\sum_1^sD_i$ be a reduced normal crossing divisor in $M$ and 

281:  $p_1,p_2,\ldots,p_s$ be positive integer numbers  greater than one.

282:  The holomorphic map $\tau:\t{M}\rightarrow M$ is called a ramification map 

283:  with divisor $D$ and ramification index  $p_i$ at $D_i$ if

284: \begin{enumerate}

285: \item

286: $\tau^{-1}(D)=\t{D}$ is a reduced normal crossing divisor;

287: \item

288:  For any point $\t{a}\in \t{M}$ and a normalizing coordinate

289:  $(x,y)\in\C^k\times \C^{n-k}$ around $a=\tau(\t{a})$, there is a normalizing 

290:  coordinate $(\t{x},\t{y})\in\C^k\times \C^{n-k}$ around $\t{a}$ such that 

291:  in these coordinates $a=(0,0)$, $\t{a}=(0,0)$ and $\tau$ is given by: 

292: \[

293: (\t{x},\t{y})\rightarrow (x,y)=((\t{x}_1^{p_{i_1}},\t{x}_2^{p_{i_2}},

294: \ldots,\t{x}_k^{p_{i_k}}),\t{y})

295: \]

296: \end{enumerate}

297: \end{defi}

298: From the definition, we can see that

299: \begin{enumerate}

300: \item

301: The critical points and values of $\tau$ are $\t{D}$ and $D$, respectively;

302: \item

303: $\tau\mid_{\t{M}\backslash \t{D}}$ is a finite covering map of some degree $p$;

304: \item

305: For any subset $I$ of $\{1,2,\ldots,s\}$, if $M_I$ is not empty then 

306: $\Pi_{i\in I}p_i$ divides $p$.

307: \end{enumerate}

308: We have also the subramification maps of $\tau$, which is stated in bellow.

309: \begin{cla}

310: \label{subncd2}

311: Keeping the notations in Definition ~\ref{ncd} and Proposition 

312: ~\ref{subncd}, for any ramification map $\tau:\t{M}\rightarrow M$ and $I$ 

313: a subset of $\{1,2,\ldots,s\}$ , the restriction of $\tau$ to $\t{M}_I$,

314:  namely $\tau_I$, is a ramification map with divisor $D_I$ and ramification 

315:  multiplicity $p_i$ at $D_i\cap M_I$, where $i\not\in I$. 

316:  Moreover, If $\tau$ is of degree $p$ then $\tau_I$ is of degree 

317:  $\frac{p}{\Pi_ {i\in I}p_i}$. 

318: \end{cla}

319: 

320: \begin{cla}

321: \label{cover2}

322: Let $\tau :\t{M}\rightarrow M$ be a ramification map of degree $p$ and with  

323: reduced normal crossing divisor $D$ and let $\pi_1(M\backslash D)$ be abelian.

324:  Then the following statement are true: 

325: \begin{enumerate}

326: \item

327: $\pi_1(\t{M})$ is abelian; 

328: \item

329: $p\pi_1(M)\subset \tau_*(\pi_1(\t{M}))$;

330: \item

331:  If $D$ is simple then $\tau_*: \pi_1(\t{M})\rightarrow \pi_1(M)$ 

332: is one to one;

333: \item

334:  If $D^*=\sum_{i=1}^{s^*}D^*_i$ is a reduced divisor in $M$ such that 

335: $D^*+D$ is a normal crossing divisor, then $\t{D^*}+\t{D}$ is also a 

336: normal 

337: crossing divisor. 

338: \end{enumerate}

339: \end{cla}

340: 

341: \pr The set $D$ is a finite union of some submanifolds of $M$ with real 

342: codimension greater than two, therefore every path in $\pi_1(M, x)$, \

343: where $x\in M\backslash D$, is homotopic to some path in 

344: $\pi_1(M\backslash D,x)$. Now the first and second statements are the direct 

345: consequences of  Proposition  ~\ref{cover}, Definition ~\ref{ncd} and the 

346: mentioned fact.

347: \\ 

348: Let $\Sigma$ be a small disk transverse to $D$ at $y\in D$ and 

349: $x\in\Sigma \cap M\backslash D$. 

350: Let also $a\in \pi_1(\t{M},\t{x})$, where $\tau(\t{x})=x$, and 

351: $\tau_*(a)$ be homotopic to zero in $M$ and $a$ do not intersect $\t{D}$. 

352: \\

353: Considering the long exact sequence ~(\ref{longler}) and the fact that 

354: $i(\tau_*(a))=0$, we conclude that $\tau_*(a)$ is homotopic to $k\delta$ 

355: in $M\backslash D$, where $k$ is an integer number and $\delta$ is a simple 

356: loop in $\Sigma$ around $y$. By Proposition ~\ref{cover}, this means 

357: that $a$ is homotopic to a closed path around $\t{y}$ in $\tau^{-1}(\Sigma)$, 

358: where $\tau(\t{y})=y$, which means that $a$ is homotopic to the point 

359: $\t{y}$ in $\t{M}$, and this proves the third statement.

360: \\

361: Let $a\in (\cap_{i\in I}D_i)\cap (\cap_{i\in I^*} D^*_{i})$, where

362:  $I\subset \{1,2,\ldots ,s\}$ and $I^*\subset \{1,2,\ldots ,s^*\}$.

363:   Choose a normalizing coordinate $(x,x^*,y)\in \C^r\times\C^{r^*}\times

364:   \C^{n-r-r^*}$

365:  around $a$ such that the components of $D$ (resp. $D^*$) through $p$ are 

366:  represented 

367:  by the coordinate $x$ (resp. $x^*$), where $r=\#I$ and $r^*=\#I^*$.

368:  Now the fourth statement is a direct consequence of  Definition ~\ref{ncd}.$\diamond$   

369:  % and let the local normalizing coordinates $(x,y)$ and $(x',y')$ around 

370:  % $a$ and $\tau(a)$, respectively, be chosen like as in the definition 

371:  % ~\ref{ncd}. Put $r=\# I$ and $r'=\# I'$. In the coordinates $(x',y')$ 

372:  % let $D'_{i_j}$ be given by $f_j(x',y')=0$, where $i_j\in I'$. $D+D'$ is 

373:  % normal crossing at $\tau(a)$.

374:  %  This means that $\frac{\partial (x', f_1,\ldots, f_{r'})}{\partial (x',y')}$ 

375:  %  has rank $r+r'$ at $\tau(a)=(0,0)$, which implies that 

376:  %  $\frac{\partial (f_1,\ldots, f_{r'})}{\partial y'}$ has rank 

377:  %  $r'$ at $0$. But $y=y'$ and so 

378:  %  $\frac{\partial (\t{f}_1,\ldots, \t{f}_{r'})}{\partial y}$ has rank 

379:  %  $s'$ at $0$ or equivalently, 

380:  %  $\frac{\partial (x, \t{f}_1,\ldots ,\t{f}_{r'})}{\partial (x,y)}$ 

381:  %  has rank $r+r'$ at $a=(0,0)$, which means that $\t{D}'+\t{D}$ is 

382:  %  normal crossing.$\diamond$

383: 

384: \subsection{ Construction of Ramification Maps}  

385: \ \ \ \ \

386: For any abelian group $G$ and a positive integer number $p$ define 

387: $G_p=G/pG$. We have the following properties with respect to $G_p$:

388: \begin{cla}

389: \label{group}

390: The following statements are true:

391: \begin{enumerate}

392: \item

393: Every morphism $f: G\rightarrow G'$ of abelian groups induces a natural 

394: morphism $f_p :G_p\rightarrow G'_p$;

395: \item

396: $(G_p)_q=G_{(p,q)}$, where $(p,q)$ denotes the greatest common divisor of 

397: $p$ and $q$;

398: \item

399: If $f:G\rightarrow G'$ is surjective then $f_p$ is also surjective. 

400: If $f$ is one to one then $f_p$ may not be one to one and so we cannot 

401: rewrite exact sequences of abelian groups by this change of groups and maps;

402: \item

403: Let $f:G\rightarrow G'$ be a morphism of abelian groups and $p$,$q$ be two

404: positive integer numbers. If $f$ is one to one, $pG'\subset f(G)$ and

405:  $(p,q)=1$ then $f_q$ is an isomorphism between $G_q$ and $G'_q$;

406: \end{enumerate}

407: \end{cla}

408: 

409: \pr We only prove the fourth statement, since others are trivial. There exist

410:  integer numbers $x,y$ such that $px+qy=1$. 

411:  \\

412: For any $a\in G'$ we have 

413: \[

414: a-q(ay)=p(ax)\in f(G)

415: \]

416:  and so $f_q$ is surjective. 

417:  \\

418: If $f_q(a)=0$ then $f(a)=qb$ for some $b\in G'$. 

419: We have 

420: \[

421: b=p(bx)+f(ay)=f(s),\ s\in G

422: \]

423:  which implies that $f(a-qs)=0$. The morphism $f$ is one to one and so 

424:  we have $a=qs$, which means that $f_q$ is one to one.$\diamond$

425: \\

426: 

427: The following statement gives us an example of ramification map with simple 

428: divisor.

429: 

430: \begin{cla}

431: Let $M$ be a complex manifold with $\pi_1(M)=0$ and $D$ be a simple divisor

432:  whose complement in $M$ has abelian fundamental group. 

433:  Let also $p'$ be a positive integer number and 

434:  \[

435:  \pi_1(M\backslash D)_{p'}=\Z_p

436:  \]

437: Then there exists a degree $p$ ramification map with divisor $D$ 

438: and ramification multiplicity $p$ at $D$.

439: \label{newman}

440: \end{cla}

441: 

442: \pr  For any $e\in D$ define $\pi_1(M\backslash D,e)=\{0_e\}$ and

443: \[

444: \tilde{M}=\cup_{e\in M} \pi_1(M\backslash D,e)_{p'}

445: \]

446: $\tilde{M}$ has the structure of a complex manifold. 

447: For any $[a_e]\in \pi_1(M\backslash D,e)$ we must define a base open set 

448: and a chart map around $[a_e]$. Consider two cases:

449: \\

450: 1. $e\in M\backslash D$

451: \\

452: Let $V_e$ be a simply connected open neighborhood of $e$ in 

453: $M\backslash D$. The following function is well-defined:

454: \[

455: \eta :V_e\rightarrow \t{M}

456: \]

457: \[

458: \eta(y)=[A_{ey}a_eA_{ey}^{-1}]

459: \]

460: where $A_{ey}$ is a path which connects $e$ to $y$ in $V_e$. The image of $\eta$ is

461:  a base open set around $a_e$ and $\eta$ is a chart map.

462: \\

463: 2. $e\in D$

464: \\

465:  Let $(V_e,(x,y))$, $(x,y)\in (\C^{n-1}\times \C ,0)$, be a coordinate around

466:   $e$ such that in this coordinate $e=(0,0)$ and $D$ is given by $y=0$. 

467:   By Leray isomorphism, for any $e'=(x',y')\in V_e$ the group 

468:   $\pi_1(M\backslash D ,e')_{p'}\simeq \Z_p$ is generated by a simple 

469:   loop around $(x',0)$ in 

470:   \[

471:   \Sigma _{(x',0)}=\{(x,y)\in (\C^{n-1}\times \C, 0)\mid x=x'\}

472:   \]

473:  In particular, we have  $p\mid p'$.

474:    This gives us the following construction of a chart map 

475:   around $[a_e]$:

476: \\

477:  For any $y_0\in (\C,0)$, let $j(y_0)$ be a point in $(\C,0)$ such that

478:  \[

479:  y_0^p=j(y_0)^p\ \&\ 0\leq arg(j(y_0))< \frac{2\pi}{p}

480:  \]

481:  and let $\delta_{y_0}$ be the path which connects $y_0$ to $j(y_0)$ in 

482:  $\{y\in (\C,0)\mid y^p=y_0^p\}$ in the clock direction.

483:  \\

484:  The image of $\delta_{y_0}$ by the map $i(y)=y^p$, $(\delta_{y_0})^p$, 

485:  is a closed path with initial and end point $y_0^p$ and so the following

486:  function is well-defined

487: \[

488: \eta :(\C^{n-1}\times\C ,0)\rightarrow \t{M}

489: \]

490: \[

491: \eta (x,y)=\{x\}\times (\delta_y)^p

492: \]

493: The image of $\eta$ is a base open set around the point $e$ and $\eta$ is a 

494: chart map. The reader can verify easily that $\t{M}$ with these base open sets

495: and chart maps is a complex manifold and

496: the natural function $\tau :\t{M}\rightarrow M$ 

497: is the desired ramification map.

498: $\diamond$       

499: 

500: \begin{theo}

501: \label{ramtheo}

502: Let $M$ be a complex manifold with $\pi_1(M)=0$ and $D=\sum_{i=1}^sD_i$ be a reduced normal 

503: crossing divisor such that  the complement of each $D_i$ in $M$ has

504:  abelian fundamental group. 

505:  Let also $p'_1,p_2',\ldots p_s'$ be  positive integer numbers which 

506:  are prime to each other. Put $p_i=\# \pi_1(M\backslash D_i)_{p_i'}$.

507:  Then there is a degree $p_1p_2\cdots p_s$ ramification map with divisor $D$

508:   and ramification multiplicity $p_i$ at $D_i$, $i=1,2,\ldots,s$.

509: \end{theo}

510: 

511: \pr The proof is by induction on $s$. For $s=1$ it is Proposition 

512: ~\ref{newman}. 

513: \\

514: Suppose that the theorem is true for $s-1$. 

515: Let $\tau :\t{M}\rightarrow M$ be a degree $p_1$ ramification map  

516: with simple divisor $D_1$ and multiplicity $p_1$ at $D_1$. 

517: We check the assumptions of the theorem for the divisor 

518: $\t{D_2}+\ldots+\t{D_s}$ in  the manifold $\t{M}$,

519:  to apply the hypothesis of the induction.

520:  \\

521: By the third part of Proposition ~\ref{cover2}

522: $\tau_* : \pi_1(\t{M})\rightarrow \pi_1(M)$ is one to one and by hypothesis

523:  $\pi_1(M)=0$, therefore $\pi_1(\t{M})=0$.

524: \\

525: Applying  Proposition ~\ref{cover2} to the ramification map 

526: $\tau\mid_{\t{M}\backslash \t{D_i}}$, we see 

527: that $\pi_1(\t{M}\backslash \t{D_i})$ is abelian;

528:  also  4 of Proposition ~\ref{cover2}  

529:  implies that $\t{D}$ is a normal crossing divisor. 

530: \\

531: The morphism 

532: \[

533: \tau _*: \pi_1(\t{M}\backslash \t{D_i})\rightarrow \pi_1(M\backslash D_i)

534: \]

535:  is one to one and by 2 of Proposition ~\ref{cover2} 

536:  \[

537:  p_1\pi_1(M\backslash D_i)\subset \tau_* (\pi_1(\t{M}\backslash \t{D_i})

538:  \]

539:  But $g.c.d.(p_1,p'_i)=1$ and so by 4 of Proposition ~\ref{group}

540:  we have

541: \[

542:      \pi_1(\t{M}\backslash \t{D_i})_{p_i'}\simeq 

543:      \pi_1(M\backslash D_i)_{p_i'}%\simeq \Z_{p_i}

544: \] 

545: Now we can apply the hypothesis of the induction to $\t{M}$ and 

546: $D'=\sum_2^sD_i$. There exists a degree $p_2\cdots p_s$ ramification map 

547: $\tau':\t{M}'\rightarrow \t{M}$ with divisor $D'$ and multiplicity $p_i$ at 

548: $D_i$, $i=2,\ldots ,s$. 

549: The reader can check that the  map $\tau\circ\tau '$ is the desired 

550: ramification map. $\diamond$

551: \subsection{Multivalued Functions}

552: \ \ \ \ \

553: To study multivalued functions, we will need to study a certain class of

554: ramification maps. 

555: First, we give the precise definition of multivalued functions.

556: \begin{defi}\rm

557: Let $\tau:\t{M}\rightarrow M$ be a degree $p$ holomorphic map between two 

558: complex manifold $\t{M}$ and $M$ i.e., $\tau$ is a finite covering map of 

559: degree $p$ out of its critical points.

560:  Every meromorphic function $g$ on $\t{M}$ is called a $p$-valued meromorphic

561:  function on $M$.

562:   Roughly speaking, the image of a point $x\in M$ under $g$ is the

563:    set $g(\tau^{-1}(x))$. The map $\tau$ is called the ramification map of $g$ 

564:    and the set of critical values of the map $\tau$ is called the ramification 

565:    divisor of the multivalued function $g$.   

566: \end{defi}

567: 

568: Given a complex manifold $M$, a meromorphic function $f$ on it and 

569: an integer number $N$. 

570: Can we  construct a ramification map of the multivalued function 

571: $f^{\frac{1}{N}}$ according to the above definition?

572:  Here we will answer to this question for some limited classes of meromorphic 

573:  functions.

574: 

575: Let $f$ be a meromorphic function on the manifold $M$ 

576: and $\tau:\t{M}\rightarrow M$ be a ramification map with reduced normal 

577: crossing divisor $D=\sum_{i=1}^sD_i$ and multiplicity $p_i$ at $D_i$, 

578: $i=1,2,\ldots,s$.

579:  Let also $div(f)=\sum m_j V_j$. Then

580: \[

581: div(\t{f})=\sum a_jm_jV_j 

582: \]

583: where $\tilde{f}=f\circ\tau$, $a_j=p_i$ if $V_j=D_i$ for some $i$ and $a_j=1$ otherwise.

584: \begin{cla}

585: Keeping the notations used above, suppose that $H^1(\t{M} ,\Z _N)=0$ and 

586: $N\mid div(\t{f})$ i.e., $N$ divides the multiplicities of the components of 

587: $div(\t{f})$, where $N$ is a positive integer number. 

588: Then $\t{f}^{\frac{1}{N}}$ is a well-defined meromorphic function on $\t{M}$.

589: Therefore, we can view $f^{\frac{1}{N}}$ as a multivalued function on $M$ with

590: the ramification map $\tau$.

591: \end{cla}

592: 

593: \pr For any point $x\in \t{M}$ there is a neighborhood $V_i$ of $x$

594:  such that in this neighborhood $\t{f}=g_i^N$, where $g_i$ is a meromorphic 

595:  function on $V_i$. Let $c_{ij}=\frac{g_i}{g_j}$, then $c_{ij}^N=1$. 

596:  Since $H^1(\t{M} ,\Z _N)=0$, there exist complex numbers $c_i$'s  

597:  such that $c_{ij}=\frac{c_i}{c_j}$. 

598:  Now $g\mid_{V_i}=\frac{g}{c_i}$'s define the global meromorphic function which is the 

599:  desired candidate for $\t{f}^{\frac{1}{N}}$. 

600:  $\diamond$    

601: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

602: \section{Picard-Lefschetz Theory}

603: \label{picsec}

604: \ \ \ \ \

605: In 1924 S. Lefschetz published his famous article \cite{lef} on the topology of 

606: algebraic varieties. In his article, in order to study the topology of an

607: algebraic variety, he considered a pencil of hyperplanes in general position 

608: with respect to that variety. 

609: Many of the Lefschetz intuitive arguments are made precise by appearance 

610: a critical fiber bundle map. In the first part of this section we introduce

611: the basic concepts of Picard-Lefschetz Theory and in the second part we 

612: introduce the Lefschetz pencil and state our two basic theorems

613: ~\ref{ggenerator},~\ref{monodromy}. This section is mainly based on the articles \cite{lam},\cite{che}. Homologies  are considered in an

614: arbitrary field of characteristic zero except it mentioned explicitly.

615: \subsection{Critical Fiber Bundle Maps} 

616: \ \ \ \ \

617: The following theorem gives us a huge number of fiber bundle  maps. 

618: \begin{Theo}

619: (Ehresmann's Fibration Theorem \cite{ehr}). 

620: Let $f:Y\rightarrow B$ be a proper submersion between the manifolds $Y$ and $B$.

621:  Then $f$ fibers $Y$ locally trivially i.e., for every point $b\in B$ 

622:  there is a neighborhood $U$ of $b$ and a $C^\infty$-diffeomorphism 

623:  $\phi :U\times f^{-1}(b) \rightarrow f^{-1}(U)$ such that 

624:  $f\circ\phi =\pi _1 =$ the first projection. Moreover if $N\subset Y$ 

625:  is a closed submanifold such that $f\mid _N$ is still a submersion then 

626:  $f$ fibers $Y$ locally trivially over N i.e., the diffeomorphism  

627:  $\phi$ above  can be chosen to carry  $U\times (f^{-1}(b)\cap N)$ 

628:  onto $f^{-1}(U)\cap N$.

629: \end{Theo}

630: 

631: The map $\phi$ is called the fiber bundle trivialization map. 

632: Ehresmann's theorem can be rewrite for manifolds with boundary and also 

633: for stratified analytic sets. In the last case the result is known as the 

634: Thom-Mather theorem.

635: 

636: In the above theorem let $f$ not be submersion, and let $C'$ be the union of 

637: critical values of $f$ and critical values of $f\mid _N$, and $C$ be the 

638: closure of $C'$ in $B$. By a critical point of the map $f$ we mean the point 

639: in which $f$ is not submersion. Now we can apply the theorem to the function 

640: \[

641: f: Y\backslash f^{-1}(C)\rightarrow B\backslash C=B'

642: \]

643: For any set $K\subset B$, we use the following notations

644: \[

645: Y_{K}=f^{-1}(K) ,\ Y'_K=Y_K\cap N,\ L_K=Y_K\backslash Y'_K

646: \]

647: and for any point $c\in B$, by $Y_c$ we mean the set $Y_{\{c\}}$. By  

648: \[

649: f:(Y,N)\rightarrow B

650: \]

651:  we mean the mentioned map and we call it the critical fiber bundle map. 

652: 

653: \begin{defi}\rm

654: Let $A\subset R\subset S$ be topological spaces. $R$ is called a strong 

655: deformation retract of $S$ over $A$ if there is a continuous map 

656: $r:[0,1]\times S\rightarrow S$ such that 

657: \begin{enumerate}

658: \item

659: $r(0,.)=id$;

660: \item

661: $r(1,x)\in R\ \&\ r(1,y)=y \ \forall x,y\in S,\ y\in R$; 

662: \item

663: $r(t,x)=x \ \forall t\in[0,1],\ x\in A$.

664: \end{enumerate}

665: Here $r$ is called the contraction map. 

666: In a similar way we can do this definition for the pairs of spaces 

667: $(R_1, R_2)\subset (S_1,S_2)$, where $R_2\subset R_1$ and $S_2\subset S_1$.

668: \end{defi}

669: 

670: We use the following important theorem to define generalized vanishing cycle 

671: and also to find  relations between the homology groups of $Y\backslash N$ 

672: and the generic fiber $L_c$ of $f$.

673: \begin{theo}

674: \label{contraction}

675: Let $f:Y\rightarrow B$ and $C'$ as before, $A\subset R\subset S\subset B$ and $S\cap C$ be a subset of the interior 

676: of $A$ in $S$, then every retraction from $S$ to $R$ over $A$ can be lifted

677:  to a retraction from $L_S$ to $L_R$ over $L_A$. 

678: \end{theo}

679: 

680: \pr According to Ehresmann's fibration theorem 

681: $f:L_{S\backslash C}\rightarrow S\backslash C$ is a $C^\infty$ locally trivial 

682: fiber bundle. The homotopy covering theorem, see 14,11.3\cite{ste}, implies 

683: that the contraction of $S\backslash C$ to $R\backslash C$ over 

684: $A\backslash C$ can be lifted so that $L_{R\backslash C}$ becomes a strong 

685: deformation retract of  $L_{S\backslash C}$ over $L_{A\backslash C}$. 

686: Since $C\cap S$ is a subset of the interior of $A$ in $S$,

687:  the singular fibers can be filled in such a way that 

688:  $L_R$ is a deformation retract of $L_S$ over $L_A$.$\diamond$

689:  \\

690:  

691:  {\bf Monodromy:} Let $\lambda$ be a path in $B'=B\backslash C$ with the initial 

692: and end points $b_0$ and $b_1$. In the sequel by $\lambda$ we will mean both 

693: the path $\lambda :[0,1]\rightarrow B$ and  the image of $\lambda$; 

694: the meaning being clear from the text.

695: \begin{cla}

696: \label{isotopy}

697: There is an isotopy 

698: \[

699: H:L_{b_0}\times[0,1]\rightarrow L_{\lambda}

700: \]

701: such that for all $x\in L_{b_0},\ t\in [0,1]$ and  $y\in N$

702: \begin{equation}

703: \label{iso}

704:  H(x,0)=x,\ H(x,t)\in L_{\lambda (t)},\  H(y,t)\in N

705: \end{equation}

706: For every $t\in [0,1]$ the map $h_t=H(.,t)$ is a homeomorphism between 

707: $L_{b_0}$ and $L_{\lambda (t)}$. The different choices of $H$

708:  and paths homotopic

709:  to $\lambda$ would give the class of homotopic maps 

710:  \[

711:  \{h_{\lambda} : L_{b_0} \rightarrow L_{b_1}\} 

712:  \]

713:  where $h_\lambda=H(.,1)$.

714:  \end{cla}

715: 

716: \pr The interval $[0,1]$ is compact and the local trivializations of

717:  $L_\lambda$ can be fitted together along $\gamma$ to yield an isotopy $H$.

718:   $\diamond$

719: 

720: The class $\{h_{\lambda} : L_{b_0} \rightarrow L_{b_1}\} $ defines the maps

721: \[

722: h_{\lambda}:\pi _*(L_{b_0})\rightarrow\pi_*(L_{b_1})

723: \]

724: \[ 

725: h _{\lambda} :H_*(L_{b_0}) \rightarrow H_*(L_{b_1})

726: \]

727: In what follows we will consider the homology class of cycles, but many of 

728: the arguments can be rewritten for their homotopy class. 

729: \begin{defi}\rm

730: For any regular value $b$ of $f$, we can define 

731: \[

732: h:\pi _1(B',b)\times H_*(L_b)\rightarrow H_*(L_{b})

733: \]

734: \[

735: h(\lambda ,.)=h_\lambda(.)

736: \]

737: $\pi_1(B',b)$ is called the monodromy group and its action 

738: $h$ on $H_*(L_b)$ is called the action of monodromy on the homology groups of

739: $L_b$.

740: \end{defi}

741: Following the article \cite{che}, we give the generalized definition of vanishing cycles.      

742: \begin{defi}

743: Let $K$ be a subset of $B$ and  $b$ be a point in $ K\backslash C$. 

744: Any relative k-cycle of $L_K$ modulo $L_b$ is called  a k-thimble above 

745: $(K,b)$ and its boundary in $L_b$ is called a vanishing $(k-1)$-cycle above $K$.

746: \end{defi}

747: 

748: Let us consider the case that we will need. Let $Y$ be a complex  compact manifold,

749:  $N$ be a submanifold of $Y$ of codimension one,

750:   $B=\S$ and $f$ be a holomorphic function.

751:  The set of critical values of $f$, $C$, is a finite set.

752:  \\ 

753: Let $c_i\in C$ (which is an 

754: isolated point of $C$ in $\S$), $D_i$ be an small 

755: disk around $c_i$ and $\tilde{\lambda _i}$ be a path in $B'$ which connects

756:  $b\in B'$ to $b_i\in \partial D_i$. 

757:  Put $\lambda _i$ the path $\tilde{\lambda _i}$ plus the path 

758:  which connects $b_i$ to $c_i$ in $D_i$ (see Figure ~\ref{alc}). Define the set $K$ in the three ways 

759:  as follows:

760: 

761: \begin{equation}

762: K^s= 

763: 	 \left\{ \begin{array}{lr}

764: 		\lambda_i & s=1\\

765: 		\lambda_i \cup D_i & s=2 \\

766: 		\tilde{\lambda}_i\cup \partial D_i & s=3\\

767: 		       \end{array} 

768: 	\right.

769: \end{equation}

770: \begin{figure}[t]

771: \begin{center}

772: \includegraphics{alc.eps}

773: \caption{}

774: \label{alc}

775: \end{center}

776: \end{figure}

777: In each case we can define the vanishing cycle in $L_b$ above $K^s$.

778: $K^1$ and $K^3$ are subsets of $K^2$ and so the vanishing cycle above $K^1$ and

779: $K^3$ is also vanishing above $K^2$. 

780: In the first case we have the intuitional concept of vanishing cycle. 

781: If $c_i$ is a critical point of $f\mid_N$ we can see that the vanishing cycle

782: above $K^2$ may not be vanishing above $K^1$.  

783: \\

784: The third case gives us the vanishing cycles obtained by a monodromy around 

785: $c_i$. In this case we have the Wang isomorphism

786: \[

787: v: H_{k-1}(L_b)\tilde{\rightarrow} H_k(L_K,L_b)

788: \]

789: Roughly speaking, The image of the cycle $\alpha$ by $v$ is the footprint

790: of $\alpha$, taking the monodromy around $c_i$. 

791: Let $\gamma_i$ be the closed path which parameterize $K_3$ i.e., 

792: $\gamma _i$ starts from $b$, goes along $\tilde{\lambda}_i$ until $b_i$, 

793: turn around $c_i$ on $\partial D_i$ and finally comes back to $b$ along 

794: $\tilde{\lambda_i}$. 

795: Let also $h_{\gamma _i}:H_k(L_{b_i}) \rightarrow H_k(L_{b_i})$ 

796: be the monodromy around the critical value $c_i$. 

797: It is easy to check that 

798: \[

799: \sigma\circ v= h_{\lambda_i}-I

800: \]

801:  where $\sigma$ is the boundary operator, 

802: therefore the cycle $\alpha$ is a vanishing cycle above $K^3$ if and 

803: only if it is in the image of $h_{\lambda_i}-I$.

804: For more information about the generalized vanishing cycle the reader is

805: referred to \cite{che}.

806: \\

807: 

808: {\bf Lefschetz Vanishing Cycle:} Let $f$ have a nondegenerate critical point

809:  $p_i$ in $Y\backslash N$ and $p_i$ be the unique critical point of 

810:  $f:(Y,N)\rightarrow \S$ within $Y_{c_i}$, where $c_i=f(p_i)$.

811:  

812: \begin{cla}

813: \label{uvan}

814: In the above situation, the following statements are true:

815: \begin{enumerate}

816: \item

817:  For all $k\not = n$ we have $H_k(L_{\lambda_i} ,L_b)=0$. This means that there 

818:  is no $(k-1)$-vanishing cycle along $\lambda_i$ for $k\not =n$;

819: \item

820: $H_n(L_{\lambda_i} ,L_b)$ is infinite cyclic generated by a hemispherical 

821: homology class $[\Delta_i]$ which is called the Lefschetz thimble and its 

822: boundary is called the Lefschetz vanishing cycle;   

823: \item

824:  Let $\lambda_i'$ be another path which connects $b$ to $c_i$ in $B'$ and is

825:   homotopic to $\lambda_i$ in $B'$ (with fixed initial and end point), 

826: then we have the same,  up to homotopy and

827:   change of sign, Lefschetz vanishing cycle in $L_b$.   

828: \end{enumerate}

829: \end{cla}

830:  For the proof of above Proposition see 5.4.1 of \cite{lam}  .

831: \\

832: By a hemispherical homology class, we mean the image of a generator of 

833: infinite cyclic group $H_n({\Bbb B}^n, {\Bbb S}^{n-1})$ under the homeomorphism 

834: induced by a continuous mapping of the closed $n$-ball ${\Bbb B}^n$ into $L_{\lambda_i}$ which sends its boundary, the $(n-1)$-sphere ${\Bbb S}^{n-1}$, to $L_{b}$. Let $B$ be a

835: small ball around $p_i$ such that in $B$ we can write $f$ in the Morse form

836: \[

837: f=c_i+x_1^2+x_2^2+\cdots+x_n^2

838: \]

839: For $b$ such that $b-c_i$ is positive real, the Lefschetz vanishing cycle 

840: in the fiber $L_b$ is given by:

841: \[

842: \delta_i =\{ (x_1,\cdots ,x_n)\in {\Bbb R}^n\mid \sum x_j^2=b-c_i\}

843: \]

844: which is the boundary of the thimble

845: \[

846: \Delta_i =\{(x_1,\cdots ,x_n)\in {\Bbb R}^n\mid \sum x_j^2\leq b-c_i\}

847: \]

848: In the above situation the monodromy $h_i$ around the critical value $c_i$ is given by the Picard-Lefschetz formula

849: \[

850: h(\delta)=\delta+(-1)^{\frac{n(n+1)}{2}}<\delta,\delta_i>\delta _i,\ \ \ \delta\in H_{n-1}(L_b)

851: \]

852: where  $<.,.>$ denotes the intersection number of two cycles in $L_{b}$.

853: \\

854: {\bf Remark:} In the above example vanishing above

855:  $K^1$ and $K^2$ are the same. Also by the Picard-Lefschetz formula the reader can verify that three types of the definition of a 

856:  vanishing cycle coincide.  

857:  In what follows by vanishing along the path $\lambda_i$ we will mean vanishing

858:  above $K^2$.

859:  %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

860: \subsection{Vanishing Cycles as Generators} 

861: \ \ \ \ \

862: Now let $\{c_1,c_2,\ldots,c_s\}$ be a subset of the set $C$ of  critical values of

863: $f$, and $b\in \S\backslash C$. 

864: Consider a system of $s$  paths $\lambda _1,\ldots,\lambda _s$ starting 

865: from $b$ and ending at $c_1, c_2,\ldots , c_s$, respectively, and such that:

866: \begin{enumerate}

867: \item

868:  each path $\lambda _i$ has no self intersection points  ; 

869: \item

870: two distinct path $\lambda _i$ and $\lambda _j$ meet only at their 

871:  common origin $\lambda _i(0)=\lambda _j(0)=b$ (see Figure ~\ref{monono1}). 

872: \end{enumerate}

873:  This system of paths is called  a distinguished system of paths. 

874:  The set of vanishing cycles along the paths 

875:  $\lambda _i,\ i=1,\ldots ,s$ is called a distinguished set of vanishing 

876:  cycles related to the critical points $c_1,c_2,\ldots,c_s$.

877: 

878: \begin{theo}

879: \label{ggenerator}

880: Suppose that $H_{k-1}(L_{\S\backslash\{a\}})=0$ for some positive integer number $k$

881:  and $a\in \S$, which may be a critical value. 

882:  Then a distinguished set of vanishing $(k-1)$-cycles 

883:  related to the critical points in the set 

884:   $C\backslash \{a\}=\{c_1,c_2,\ldots,c_r\}$ generates $H_{k-1}(L_b)$.

885: \end{theo}

886: 

887: \pr We use the arguments of the article \cite{lam} Section 5. Note that in our case the fiber is $L_b=Y_b\backslash N$ and not $Y_b$.

888: \\

889: We consider our system of distinguished paths inside a large disk 

890: $D_+$ so that $a\in \S\backslash \overline{D}_+$, the point $b$ 

891: is in the  boundary of $D_+$ and all critical values $c_i$'s in 

892: $C\backslash \{a\}$ are interior points of $D_+$. 

893: Small disks $D_i$ with centers $c_i$ $i=1,\cdots ,r$ are chosen so that they

894:  are mutually disjoint and contained in $D_+$. 

895:  Put 

896:  \[

897:  K_i=\lambda _i\cup D_i,\ K=\cup_{i=1}^r K_i

898:  \]

899:    The pair $(K,b)$ is a strong deformation retract of $(D_+,b)$ 

900:    and so by Theorem ~\ref{contraction} $(L_K,L_b)$ is a strong deformation retract of

901:     $(L_{D_+},L_b)$. The set $\tilde{\lambda}=\cup \tilde{\lambda_i}$ 

902:     can be retract within itself to the point $b$ and so $(L_K,L_b)$ and 

903:     $(L_K,L_{\tilde{\lambda}})$ have the same homotopy type. 

904:     By the excision theorem (see \cite{mas}) we conclude that 

905: \[

906: H_k(L_{D_+},L_{b})\simeq\sum _{i=1}^{r} H_k(L_{K_i},L_{b})\simeq\sum _{i=1}^{r} H_k(L_{D_i},L_{b_i})

907: \]

908: \begin{figure}[t]

909: \begin{center}

910: \includegraphics{mono1.eps}

911: \caption{}

912: \label{monono1}

913: \end{center}

914: \end{figure}

915: Write the long exact sequence of the pair $(L_{D_+},L_b)$:

916: \begin{equation}

917: \ldots\rightarrow H_k(L_{D_+})

918: \rightarrow H_k(L_{D_+},L_b) \stackrel{\sigma}{\rightarrow } H_{k-1}(L_b) 

919: \rightarrow H_{k-1}(L_{D_+})\rightarrow \ldots

920: \label{long}

921: \end{equation}

922: Knowing this long exact sequence, 

923: it is enough to prove that $H_{k-1}(L_{D_+})=0$. 

924: A contraction from $\S\backslash \{a\}$ to $D_+$ can be lifted

925:  to the contraction of $L_{\S\backslash \{a\}}$ to $L_{D_+}$ 

926:  which means that $L_{D_+}$ and $L_{\S\backslash\{a\}}$ have the same 

927:  homotopy type and so by the  hypothesis $H_{k-1}(L_{D_+})=0$.$\diamond$ 

928: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

929: \subsection{Lefschetz Pencil}

930: \ \ \ \ \

931: In this section we repeat some notations and propositions  of \cite{lam} Section

932: 2. All the proofs can be found there.  

933: \\   

934: The hyperplanes of $\pn$ are  points of the dual projective space $\pnh$. 

935: We use the following notation:

936: \[

937: H_y \subset \pn\ , \ y\in\pnh

938: \]

939: Let $X$ be a closed irreducible subvariety of $\pn$ and let $X_e \subset X$ 

940: be the nonempty open subset of its regular points. Define 

941: \[

942:      V_X^{'}=\{(x,y)\in \pn \times \pnh \mid x\in X_e\ \&\ H_y 

943:      \hbox{ is tangent to

944:      $X$ at $x$}\}

945: \]

946: This is a quasiprojective subset of $\pn \times \pnh$, because the set

947: \[

948: 	\tilde{V}=\{(x,y)\in \pn \times \pnh \mid 

949: 	\hbox{ $x$ is a singular point of $X$

950:         or $H_y$ is tangent to}

951: \]

952: \[

953:  	X \hbox{ at } x \}

954: \]

955: is closed in $\pn \times \pnh $ and $V_X^{'}$ is a Zariski open in 

956: $\tilde{V}$. The closure $V_X$ of $V_X^{'}$ is called the tangent hyperplane 

957: bundle of $X$. Consider the second projection 

958: \[

959: 	\pi _2 : V_X \rightarrow \pnh \ ,\ (x,y)\rightarrow y

960: \]

961: its image $\check{X}$ is a closed irreducible subvariety of $\pnh$ of dimension 

962: at most $n-1$ which is called the dual variety of $X$.

963: If $X$ is a smooth variety then

964: \[

965:  \check{X}=\{y\in \pnh \mid H_y \hbox{ is tangent to $X$ at some point }\}

966: \]

967: In general $\check{X}$

968:  has singularities even if $X$ does not. 

969:  If $dim(\check{X})=N-1$ the degree of $\check{X}$ is well-defined and if 

970:  $dim(\check{X})<N-1$ we define $deg(\check{X})=0$.  

971: \begin{cla}

972: (Duality Theorem \cite{lam} 2.2) The tangent hyperplane bundles of $X$ and $\check{X}$ coincide

973: \[

974: V_X=V_{\check{X}} \hbox{ and hence } \check{\check{X}}=X

975: \]

976: \end{cla}

977: 

978: A pencil in $\pn$ consists of  all hyperplanes which contain a fixed 

979: $(N-2)$-dimensional projective space  $A$, which is called the axis 

980: of the pencil. We denote a pencil by 

981: \[

982: \{H_t\}_{t\in G}

983: \]

984: or $G$ itself, where $G$ is a projective line in $\pnh$. 

985: \\

986: The pencil $\{H_t\}_{t\in G}$ is in general position with respect to $X$

987:  if $G$ is in general position with respect to $\check{X}$. 

988:   From now on, fix a  pencil $\{H_t\}_{t\in G}$ 

989:   in general position with

990:   respect to $X$.

991:   

992: \begin{cla}

993: ( \cite{lam}, 1.6.1) The axis A intersects X transversally.

994: \end{cla}

995: For the pencil $\{H_t\}_{t\in G}$ put 

996: \[

997: X_t=X\cap H_t,\ L_t=X_t\backslash A ,\ C=G\cap \check{X}=

998: \{c_1,c_2,\ldots,c_r\},\ X'=A\cap X  

999: \]

1000: We will sometimes parameterize $G$ by $\S$ and denote the pencil by 

1001:   $\{H_t\}_{t\in \S}$.

1002: In order to have a map whose level surfaces are the $X_t$'s,  we need to do a "blow up"

1003:  along the  variety $X'$. Let 

1004: \[

1005: Y=\{(x,t)\in X\times \S \mid x\in H_t\}

1006: \]

1007: There are two projections

1008: \[

1009: X\stackrel{p}{\leftarrow} Y \stackrel{f}{\rightarrow} \S

1010: \]

1011: Put $Y'=p^{-1}(X')=X'\times \S$ then

1012: \begin{cla}

1013: \label{fibration}

1014: (\cite{lam} 1.6.2, 1.6.3, 1.6.4) If $X$ is a smooth variety then

1015: \begin{enumerate}

1016: \item

1017: The modification $Y$ of $X$ is smooth and irreducible;

1018: \item 

1019: $p$ is an isomorphism between $Y\backslash Y'$ and $X\backslash X'$ and also 

1020: an isomorphism between $f^{-1}(t)$ and $X_t$;

1021: \item

1022: For every critical value $c_i,\ i=1,\ldots ,r$  of $f$, the hyperplane 

1023:  $H_{c_i}$ has a unique tangency of order two with $X$ which lies out of $A$.

1024:  The other hyperplanes $H_c,\ c\not\in C$ are transverse to $X$;

1025: \item

1026: The projection $f:Y\rightarrow \S$ has $r=deg(\check{X})$ nondegenerate 

1027: critical points $p_1,\ldots, p_r$ in $Y\backslash Y'$ such that $f(p_i)=c_i$'s 

1028: are distinct 

1029: values in $\S$.

1030: \end{enumerate}

1031:  \end{cla}

1032: 

1033: Now we have the critical fiber bundle map 

1034: $f:(Y, Y')\rightarrow \S$. Note that $f\mid _{Y'}$ has no critical points. 

1035:  We conclude that the natural function $f:X\backslash A\rightarrow \S$ is a

1036:  fiber bundle map over $\S \backslash C$.

1037: 

1038: \begin{defi}\rm 

1039: We can view $f:X\rightarrow \S$ as a meromorphic function on $X$. $f$ is called

1040: the Lefschetz meromorphic function. The foliation induced by the pencil

1041: $\{H_t\}_{t\in \S}$ is called the Lefschetz foliation.

1042: \end{defi}

1043: 

1044: \begin{cla}

1045: \label{equiv}

1046: The pencil $\{H_t\}_{t\in G}$ is in general position with respect to $X$ if

1047:  and only if

1048:  \begin{enumerate}

1049:  \item

1050:  Choosing a good parameterization of $G$, $(f)_\infty=f^{-1}(\infty)$ and 

1051: $(f)_0=f^{-1}(0)$ are smooth irreducible varieties and intersect each

1052:  other transversally and ;

1053: \item

1054: $f$ has only nondegenerate critical points with distinct images.

1055: \end{enumerate}

1056:  \end{cla}

1057: 

1058: \pr If $\{H_t\}_{t\in G}$ is in general position with respect to $X$, then the axis

1059: of the pencil, $A$, intersects $X$ transversally. Knowing that

1060: \[

1061: (f)_0=H_0\cap X,\ (f)_\infty=H_\infty\cap X, \ H_0\cap H_\infty=A

1062: \]

1063: we conclude that $(f)_0$ and $(f)_\infty$ intersect each other transversally.

1064: The second statement is Proposition ~\ref{fibration}, 4.

1065: \\

1066: Now suppose that $f$ satisfies 1 and 2. Suppose that $G$ is not

1067: in general position with respect to $\check{X}$ at $x\in G\cap\check{X}$.   

1068: We can distinguish two cases

1069: \\

1070: 1. $x$ is a smooth point of $\check{X}$ and $G$ is tangent to $\check{X}$ at

1071: $x$;

1072: \\

1073: Let $H_s$ be a hyperplane which passes through $x$, contains $G$ and is tangent

1074: to $\check{X}$ at $x$. By Duality Theorem the information

1075: \[

1076: x\in G\subset H_s, \ \hbox{ $H_s$ is tangent to $\check{X}$ at $x$}

1077: \]

1078: can be translated to 

1079: \[

1080: s\in A\subset H_x, \ \hbox{ $H_x$ is tangent to $X$ at $s$}

1081: \]

1082: But this contradicts the first statement.

1083: \\

1084: 2. $x$ is a singular point of $\check{X}$;

1085: \\

1086: This case also cannot happen. By the argument used in the proof of 1.6.4 of \cite{lam}, we have: $x$ is a smooth point of $\check{X}$, if and only if, $H_x$ has a unique tangency point of order two with $X$.$\diamond$

1087: %Consider two hyperplanes $H_a$ and $H_b$ passing through $X$ and tangent to

1088: %$\check{X}$ at this point. By Duality theorem, the hyperplane $H_x$ is tangent

1089: %to $X$ at $a$ and $b$ which contradicts the second statement.$\diamond$

1090: \begin{theo}

1091: Suppose that the pencil $\{H_t\}_{t\in G }$ is in general position with respect to $X$ 

1092: and Let $a$ be a point in $\S\backslash C$. Then for every $b\in\S\backslash C$ \label{generator}

1093: \begin{enumerate}

1094: \item

1095:  $H_i(L_b)\simeq H_i(X\backslash H_a),\ i\not = n, n-1$ 

1096:   \item

1097: If $H_{n-1}(X\backslash H_a)=0$, then a distinguished set of vanishing 

1098: cycles related to the critical values $c_1,c_2,\ldots,c_r$ generates the 

1099: group $H_{n-1}(L_b)$. 

1100: \end{enumerate}

1101: \end{theo}

1102: 

1103: \pr This is a direct consequence of Theorem ~\ref{ggenerator} and  

1104: Theorem ~\ref{uvan} and the long exact sequence ~(\ref{long}).$\diamond$

1105: \\

1106: 

1107: {\bf Blow up:} Fix the point $b\in\pro$. All lines through $b$ in $\pro$ 

1108: form a projective space of dimension $n-1$, namely $P$. Define

1109: \[

1110: \t{\pro}=\{(x,y)\in \pn\times P\mid x\in y\}

1111: \]

1112: $\t{\pro}$ is a smooth subvariety of $\pn\times P$. We have two natural

1113:  projections

1114: \[

1115: \pro\stackrel{i}{\leftarrow} \t{\pro} \stackrel{f}{\rightarrow} P

1116: \]

1117: The reader can check that $f$ is an isomorphism between $i^{-1}(b)$ and $P$ 

1118: and $i$ is an isomorphism between 

1119: $\pro\backslash\{b\}$ and $\t{\pro}\backslash i^{-1}(b)$. $\t{\pro}$ is 

1120: called the blow up of $\pro$ at the point $b$.

1121:  Roughly speaking, we delete the point $b$ from $\pro$ and 

1122:  substitute it by a projective space of dimension $n-1$.

1123: 

1124: \begin{theo}

1125: \label{monodromy}

1126: Suppose that the pencil $\{H_t\}_{t\in \S}$ is in general position with respect to $X$

1127: and let $b\in\S\backslash C$, then

1128: \begin{enumerate}

1129: \item

1130: For every two Lefschetz vanishing cycles $\delta_0$ and $\delta_1$ in 

1131: $X_b$ there exists a closed path $\lambda$ in $\S\backslash C$ with initial 

1132: and end point $b$ and such that

1133: \[

1134: h_{\lambda}(\delta_0)=\pm \delta_1

1135: \]

1136: where $h_\lambda$ is the monodromy along the path $\lambda$;

1137: \item

1138:  If $H_{n-1}(X\backslash H_a)=0$ for some $a\in \S\backslash C$ and $H_{n-1}(X_b)\not =0$ 

1139:  then for every Lefschetz vanishing cycle $\delta$ in $L_b$, 

1140:  the action of the monodromy group on $\delta$ generates $H_{n-1}(L_b)$.

1141: \end{enumerate}

1142: \end{theo}

1143: 

1144: \pr 

1145: %The second statement is a direct consequence of the first and of 

1146: %Theorem ~\ref{generator}, therefore it is enough to prove the first 

1147: %statement. 

1148: The first statement and its proof can be found in 7.3.5 of \cite{lam}. But we can give a rather short proof for it as follows:  

1149: \\

1150: Let us consider the pencil $\{H_t\}_{t\in G}$ as the  projective line  $G$ 

1151: in $\check{\pro}$. Let $\delta _0$ and $\delta_1$ vanish along the paths $\lambda_0$ and 

1152: $\lambda_1$ which connect $b$ to critical values $c_0$ and $c_1$ in $G$, 

1153: respectively. The subset $Z\subset \check{X}$ consisting of all points $x$

1154:  such that the line through $x$ and $b$ is not in general position with 

1155:  respect to $\check{X}$ is a proper and algebraic subset of $\check{X}$. Since $\check{X}$ is an

1156:  irreducible variety and $c_0,c_1\in \check{X}\backslash Z$, there is a path 

1157:  $w$ in $\check{X}\backslash Z$ from $c_0$ to $c_1$. Denote by $G_s$ the line 

1158:  through $b$ and $w(s)$. After Blow up at the point $b$ and using the 

1159:  Ehresmann's theorem, we conclude that: 

1160:  \\

1161: There is an isotopy

1162: \[

1163: H:[0,1]\times G\rightarrow \cup_s G_s 

1164: \]

1165: such that  

1166: \begin{enumerate}

1167: \item

1168:  $H(0,.)$ is the identity map; 

1169:  \item 

1170:  for all $s\in [0,1]$, $H(s,.)$ is a $C^{\infty}$ isomorphism between $G$ 

1171:  and $G_s$ which sends points of $\check{X}$ to $\check{X}$ 

1172:  \item

1173:  For all $s\in [0,1]$  $H(s,b)=b$ and $H(s,c_1)=w(s)$

1174:  \end{enumerate}

1175: Let $\lambda_s'=H(s,\lambda_0)$. 

1176: In each Lefschetz pencil $\{H_t\}_{t\in G_s}$ the cycle $\delta_0$ in $X_b$ 

1177: vanishes along the path $\lambda_s'$ in $w(s)$, 

1178: therefore $\delta_0$ vanishes along $\lambda_1'$ in $c_1=w(1)$. 

1179: Consider $\lambda_1$ and $\lambda_1'$ as the paths which start from $b$ and end

1180:  in a point $b_1$ near $c_1$ and put $\lambda=\lambda _1'-\lambda_1$. 

1181:  By uniqueness of the Lefschetz vanishing cycle along a fixed path we

1182:  can see that the path $\lambda$ is the desired path. 

1183: \\

1184: Now let us prove the second part. Unfortunately the above argument is true for the fiber $X_b$ and not  $L_b$. Therefore by Theorem ~\ref{generator} we can

1185: only conclude that the action of the monodromy on a  vanishing cycle generates $H_{n-1}(X_b)$. Since $H_{n-1}(X_b)\not =0$, there is no 

1186: homologous to zero  vanishing cycle in $X_b$ . Let us prove that the intersection matrix $[<\delta_i,\delta_j>]_{r\times r}$ of vanishing cycles is connected i.e., for any two vanishing cycles $\delta$ and $\delta'$ there exists a chain $\delta_{i_1},\delta_{i_2},\ldots,\delta_{i_e}$ of vanishing cycles with the following properties: 

1187: \[

1188: \delta=\delta_{i_1}, \delta'=\delta_{i_e}\ \ \ <\delta_{i_k},\delta_{i_{k+1}}>\not=0,\ \ \ k=1,2,\ldots,e-1

1189: \]

1190: If $\delta'$ is not connected to $\delta$ as above then by Picard-Lefschetz formula $\delta'$ has intersection zero with all cycles obtained by the action of the monodromy on $\delta$. But the action of the monodromy on $\delta$ generates $H_{n-1}(X_b)$. $X_b$ is compact and so $\delta'=0$ in $X_b$ which is a contradiction.  Now using Picard-Lefschetz formula in $L_b$ we see that the action of the monodromy on a vanishing cycle generates any other vanishing cycle in $L_b$. By Theorem ~\ref{generator}, vanishing cycles generate $H_{n-1}(L_b)$ and so the second statement is proved.  $\square$ 

1191:  %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%Toplogy of Integrables

1192: \section{Topology of Integrable Foliations}

1193: \label{top}

1194: \ \ \ \ \ 

1195: In this section we will combine the results of the sections ~\ref{ramsec} and

1196: ~\ref{picsec} to generalize Theorem ~\ref{generator} and Theorem

1197: ~\ref{monodromy} for the foliation $\F(\INT)$. Note that the first Integral of

1198: $\F$ has the critical fibers $\{F=0\}$ and $\{G=0\}$, if $p>1$ and $q>1$

1199: respectively, which don't appear in the Lefschetz foliation. Homologies are

1200: considered in an arbitrary field except in the mentioned cases.   

1201: \subsection{Integrable Foliations and Lefschetz Pencil}

1202: \ \ \ \ \

1203: Let $\F(\INT)$ be an integrable foliation satisfying the generic conditions

1204:  of Proposition ~\ref{gen}. Put 

1205: \[ 

1206: D_1=\{F=0\},\ D_2=\{G=0\}, L_b=(\integ)^{-1}(b)\backslash \R,\ X_b=L_b\cup \R

1207: \]

1208: Consider the reduced normal crossing divisor $D=D_1+D_2$ 

1209: and the positive integer numbers $q,p$ such that 

1210: \[

1211: deg(F)=qd,\ deg(G)=pd,\ g.c.d.(p,q)=1

1212: \]

1213: It is a well-known fact that the fundamental group of the complement of any smooth hypersurface 

1214: $V$ in $\pro$ is isomorphic to $\Z _{deg(V)}$ and therefore

1215: \[

1216: \pi_1(\pro\backslash D_1)_q=\Z _q\  , \ \pi_1(\pro\backslash D_2)_p=\Z _p

1217: \]

1218: By Theorem ~\ref{ramtheo} there exists a degree $pq$ ramification map 

1219: \begin{equation}

1220: \label{rammap}

1221: \tau :\t{\pro}\rightarrow \pro

1222: \end{equation}

1223: with divisor $D$ and ramification

1224: multiplicities

1225:  $q$ and $p$ in $D_1$ and $D_2$, respectively. We can view the polynomials 

1226:  $F$, $G$ and the coordinates $x_i, i=0,\ldots, n-1$, as meromorphic functions 

1227:  with the pole divisor 

1228:  $H_\infty$, the hyperplane at infinity. 

1229:  We have

1230: \[

1231: div(\t{F})=q(\t{D_1} -d.\t{H}_\infty)

1232: \]

1233:  \[

1234: div(\t{G})=p(\t{D_2} -d.\t{H}_\infty)

1235: \]

1236: therefore $\t{F}^{\frac{1}{q}}$ and $\t{G}^{\frac{1}{p}}$ are well-defined 

1237: meromorphic functions on $\t{\pro}$. Define

1238: \[

1239: j:\tpro\backslash \t{H}_\infty\rightarrow \C^2

1240: \]

1241: \[

1242: j(x)=(\t{F}^{\frac{1}{q}}, \t{G}^{\frac{1}{p}})

1243: \]

1244: 

1245: The following proposition shows that the

1246:  different sheets of $\t{\pro}$ are due to the different values of 

1247:  $\t{F}^{\frac{1}{q}}$ and $\t{G}^{\frac{1}{p}}$.

1248: \begin{cla}

1249: \label{onetoone}

1250: For any $x\in \pro\backslash H_\infty$ the map $j$ 

1251: takes distinct values in $\tau^{-1}(x)$. (If $x\in H_\infty$ choose 

1252: another hyperplane as the hyperplane at infinity). 

1253: \end{cla}

1254: 

1255: \pr The set 

1256: \[

1257: S=\{x\in \pro \mid \exists a,b\in \t{\pro} \ s.t.

1258: \ \tau (a)=\tau(b)=x,\ a\not =b,\ j(a)=j(b) \}

1259: \]

1260:  is an open closed subset of $\pro$, because the values of 

1261:  $\t{F}^{\frac{1}{q}}$  ( $\t{G}^{\frac{1}{p}}$) in $\tau ^{-1}(x)$ are the same 

1262:  up to multiplication by some $q$-th ($p$-th) root of the unity. 

1263:  Choosing  normalizing coordinates like in Definition ~\ref{ncd} around the  

1264:  points $a\in D_1\cap D_2$ and $\tau^{-1}(a)$, we have 

1265: \begin{equation}

1266: (x_1, x_2)\stackrel{j}{\leftarrow}(x_1,x_2,y)\stackrel{\tau}{\rightarrow} (x_1^q,x_2^p,y) 

1267: \end{equation}

1268: $\tau$ has the degree $pq$ and so $S$ has not any point near $a$, 

1269: therefore $S$ is empty. $\diamond$

1270: \\

1271: 

1272:  The foliation $\t{\F}=\tau^{*}(\F)$ in $\t{\pro}$ is also integrable and has

1273:   the first integral $\integr$ with divisor

1274: \[

1275: div(\integr)=\t{D_1}-\t{D_2}

1276: \]

1277: For every $\t{b}\in \S$, let 

1278: \[

1279: \t{L}_{\t{b}}=(\integr)^{-1}(\t{b})\backslash

1280: \t{{\cal R}},\  \t{X}_{\t{b}}=\t{L}_{\t{b}}\cup \t{{\cal R}} ,\  \t{{\cal R}}=\t{D}_1\cap\t{D}_2

1281: \]

1282: The following proposition states the relations between the leaves of $\F$ 

1283: and $\t{\F}$. 

1284: \begin{cla}

1285: \label{a}

1286: The following statements are true:

1287: \begin{enumerate}

1288: \item

1289: $\tau$ maps $\t{\R}$ to $\R$ biholomorphically;

1290: \item

1291: $\tau\mid_{\t{L}_0}:\t{L}_0\rightarrow L_0$

1292: ($\tau\mid_{\t{L}_\infty}:\t{L}_\infty\rightarrow L_\infty$) 

1293: is a finite covering map of degree $q$ (repectively $p$);

1294: \item

1295:  For any $c\not =0,\infty$, $\tau$ maps $\t{L_c}$ to $L_{c^{pq}}$ biholomorphically.

1296: \end{enumerate}

1297: \end{cla}

1298: 

1299: \pr The first and second statements are the results obtained in Proposition 

1300: ~\ref{subncd2}.

1301:  For the third it is enough to prove that $\tau\mid_{\t{L_c}}$ is one to one.   

1302: \\

1303: If $\tau(x)=\tau(y)$ and $\integr(x)=\integr(y)$ then 

1304: $\frac{\t{F}^{\frac{1}{q}}(x)}{\t{F}^{\frac{1}{q}}(y)}=

1305: \frac{\t{G}^{\frac{1}{p}}(x)}{\t{G}^{\frac{1}{p}}(y)}$

1306: is a constant which is $p$-th and $q$-th root of the unity, but $g.c.d.(p,q)=1$ 

1307: and so  $\t{F}^{\frac{1}{q}}(x)=\t{F}^{\frac{1}{q}}(y)$ and 

1308: $\t{G}^{\frac{1}{p}}(x)=\t{G}^{\frac{1}{p}}(x)$. By Proposition 

1309: ~\ref{onetoone} we conclude that $x=y$. $\diamond$ 

1310: \\

1311: Define 

1312: \[

1313: v:\t{\pro}\rightarrow \pn

1314: \]

1315: \[

1316: v(A)=[\ldots; x_0^{i_0}x_1^{i_1}\cdots x_n^{i_n};\ldots;x_n^d;\t{F}^\frac{1}{q}(A);\t{G}^\frac{1}{p}(A)],\ i_0+\cdots +i_n=d 

1317: \]

1318: $N-2$ is the number of monomials of degree $d$ and the 

1319: variables $x_0,x_1,\ldots,x_n$.

1320: \begin{cla}

1321: \label{embedding}

1322: $v$ is an embedding.

1323: \end{cla}

1324: 

1325: \pr Consider the following commutative diagram:

1326: \begin{equation}

1327: \begin{array}{ccc}

1328: \label{di1}

1329: \pro & \stackrel{v_d}{\rightarrow} & \C P(N-2) \\

1330: \tau\uparrow &  & i\uparrow  \\

1331: \t{\pro} & \stackrel{v}{\rightarrow} & \pn 

1332: \end{array}

1333: \end{equation}

1334: where $v_d$ is the well-known veronese embedding and $i$ is the projection 

1335: on the first $N-1$ coordinates. 

1336: \\

1337: 1. $v$ is one to one;

1338: \\

1339: If $a,b\in \t{\pn}$ and $v(a)=v(b)$ then $v_d(\tau(a))=v_d(\tau(b))$ and 

1340: so $\tau(a)=\tau(b)$ and by Proposition ~\ref{onetoone} we conclude 

1341: that $a=b$.

1342: \\ 

1343: 2. $v$ is locally embedding;

1344: \\

1345: For any $a\in \t{\pro}$ choose normalizing coordinates around $a$ and $\tau(a)$. 

1346: For example , if $a\in D_1\cap D_2$ the diagram  ~(\ref{di1}) has the form 

1347: \begin{equation}

1348: \begin{array}{ccccc}

1349: \label{di2}

1350: (x_1^q,x_2^p,y) & \stackrel{v_d}{\rightarrow} & v_d(x_1^q,x_2^p,y) \\

1351: \uparrow &  & \uparrow \\

1352: (x_1,x_2,y) & \stackrel{v}{\rightarrow} & (v_d(x_1^q,x_2^p,y),x_1,x_2)

1353: \end{array}

1354: \end{equation}

1355: we have to prove that the bottom map is an embedding at $0$.

1356: \[

1357: Dv(0)=

1358: 	 \left [ \begin{array}{ccc}	

1359: 		 %\frac{\partial v_d}{\partial x_1}k & \frac{\partial

1360: 		%v_d}{\partial x_2}k'  

1361: 	* & * 	& \frac{\partial v_d}{\partial y} \\

1362: 		1 & 0 & 0 \\

1363: 		0 & 1 & 0 	

1364:        \end{array} 

1365: 	\right ].

1366: \]

1367: %where $k,k'=0,1$ depending on $q,p=1$ or $q,p\not =1$, repectively.

1368: $v_d$ is the veronese embedding and so $Dv(0)$ has the maximal rank rank $n$. 

1369: For  other points the proof is similar. $\diamond$       

1370: 

1371: The foliation $\t{\F}$ is obtained by hyperplane sections of the following Lefschetz pencil

1372: \[

1373:  \{H_t\}_{t\in \S},\ H_t=\{[x;x_{N};x_{N+1}]\in\pn\mid x_{N}=tx_{N+1}\}

1374: \]

1375: $\t{D}_1$ and $\t{D}_2$ intersect each other transversally in $\t{\cal R}=

1376: \t{D}_1\cap \t{D}_2$, and $\integr$ has nondegenerate critical points with

1377: distinct images,  therefore $\{H_t\}_{t\in \S}$ is in general 

1378: position with respect to $X=v(\t{\pro})$.

1379: 

1380: Now consider the following commutative diagram

1381: \begin{equation}

1382: \label{di3}

1383: \begin{array}{ccc}

1384: \t{\pn}& \stackrel{\tau}{\rightarrow} & \pro \\

1385: \integr\downarrow &  & \integ\downarrow \\

1386: \S & \stackrel{i}{\rightarrow} & \S

1387: \end{array}

1388: \end{equation}

1389: where $i(z)=z^{pq}$. Let $\t{C}$ denote the set of critical values of

1390: $\integr$, then by Proposition ~\ref{a}, we conclude that

1391: 

1392: \begin{coro}

1393: $\integ$ and $\integr$ are fiber bundle maps over 

1394: $\S\backslash (C\cup A)$ and $\S\backslash \t{C}$, respectively.

1395: \end{coro}

1396: 

1397: 

1398: \begin{coro}

1399: \label{kufteh}

1400: Let $b\in \S$ be a regular value of $\integ$. Then for every two Lefschetz vanishing cycles $\delta_1$ and $\delta_2$ in 

1401: $X_b$ there is a

1402:   monodromy $h_\lambda$ such that

1403: \[

1404: h_\lambda (\delta_1)=\pm \delta _2

1405: \]

1406: \end{coro}

1407: 

1408: \pr Fix a point $\t{b}\in i^{-1}(b)$.

1409:  By diagram ~\ref{di3}, we have the following commutative diagram

1410: \begin{equation}

1411: \label{dimo}

1412: \begin{array}{ccccc}

1413: \pi_1(\S\backslash \t{C},\t{b}) & \times & H_*(\t{X}_{\t{b}}) & 

1414: \rightarrow & H_*(\t{X}_{\t{b}}) \\ 

1415: i_*\downarrow & & \tau_*\downarrow & & \tau_*\downarrow \\

1416: \pi_1(\S\backslash (C\cup A),b) & \times & H_*(X_b) & \rightarrow & 

1417: H_*(X_b)

1418: \end{array}

1419: \end{equation}

1420: $\t{\delta}_i=\tau^*(\delta_i),\ i=1,2$ are two Lefschetz vanishing cycles in

1421: $\t{L}_{\t{b}}$. By Theorem ~\ref{monodromy}, there exists a path $\t{\lambda}\in

1422: \pi_1(\S\backslash \t{C},\t{b})$ such that the related monodromy takes 

1423: $\t{\delta}_1$ to $\pm \t{\delta}_2$. We can assume that this path doesn't pass

1424: through $0$ and $\infty$. Now by the above diagram the path

1425: $i(\t{\lambda})$ is the desired path.$\diamond$   

1426: 

1427: 

1428: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%5

1429: \subsection{More About the  Topology of Integrable Foliation}

1430: \def\intag{\frac{\t{F}}{\t{G}^\frac{q}{p}}}

1431: \ \ \ \ \

1432: Here we want to prove a theorem similar to Theorem ~\ref{ggenerator} for the

1433: foliation $\F(\INT)$.

1434: \\

1435: Let $\tau: \t{\pro}\rightarrow \pro$ be a ramification map with simple 

1436: divisor $D=\{G=0\}$ and multiplicity $p$ at $D$. 

1437:  \[

1438: div(\t{G})=p(\t{D} -d.\t{H}_\infty)

1439: \]

1440: therefore $\t{G}^\frac{1}{p}$ is a well-defined meromorphic function on 

1441: $\t{\pro}$. We denote by $\t{C}$ the set of critical values of $\intag$ in

1442: $\S\backslash\{\infty\}$. Also 

1443: \[

1444: \t{{\cal R}}=\{\t{F}=0\}\cap \{\t{G}=0\}

1445: \] 

1446: The foliation $\t{\F}=\tau^*(\F)$ has the first integral 

1447: $\intag$. Note that $0\in\S$ is no more a critical point of $\intag$. 

1448: Consider the following commutative diagram

1449: \[

1450: \begin{array}{ccc}

1451: \label{di3}

1452: \tpro& \stackrel{\tau}{\rightarrow} & \pro \\

1453: \intag\downarrow &  & \integ\downarrow \\

1454: \S & \stackrel{i}{\rightarrow} & \S

1455: \end{array}

1456: \]

1457: where $i(z)=z^{p}$. Like before we have

1458: 

1459: \begin{cla}

1460: The following statements are true

1461: \begin{enumerate}

1462: \item

1463: $\intag$ is a fiber bundle map over $\S\backslash (\t{C}\cup\{\infty\})$.

1464: \item

1465: $\tau$ maps $\t{\R}$ ($\t{L}_\infty$) to $\R$ (respectively $L_\infty$) 

1466: biholomorphically;

1467: \item

1468: $\tau\mid_ {\t{L}_0}:\t{L}_0\rightarrow L_0$  

1469: is a finite covering map of degree $q$;

1470: \item

1471:  For any $c\not =0,\infty$, $\tau$ maps $\t{L_c}$ to $L_{c^{p}}$ 

1472:  biholomorphically.

1473: \end{enumerate}

1474: \end{cla}

1475: 

1476: \begin{theo}

1477: \label{16apr}

1478: If $n=2$ then a distinguished set of 

1479: Lefschetz vanishing cycles related to the critical points in the set

1480: $\t{C}$ generates the first homology group of a regular fiber $\tilde{L}_b$ of $\intag$.

1481: \end{theo}

1482: 

1483: \pr (n=2) By Theorem ~\ref{generator} it is enough to prove that 

1484: $H_{n-1}(\t{\pro}\backslash \t{D})=0$.

1485: According to Proposition ~\ref{cover},

1486: $\tau_*:H_{n-1}(\tpro \backslash \t{D},\Z)\rightarrow H_{n-1}(\pro\backslash

1487: D,\Z)$ 

1488: is one to one, and we also know  that 

1489: $H_{n-1}(\pro\backslash D, \Z)=\Z_{deg(G)}$, which implies that $H_{n-1}(\pro\backslash D)=0$ in an arbitrary field. These facts imply what we want.

1490: $\diamond$  

1491: \begin{coro}

1492: \label{21agu}

1493: Let $b$ be a regular value of $\integ$ and $\Delta$ be a set of distinguished

1494: Lefschetz vanishing cycles related to the critical points in the set $C$. Let

1495: also $h$ be the monodromy around the critical value $0$. Then the set  

1496: \[

1497: \Delta\cup h(\Delta)\cup\cdots\cup h^{p-1}(\Delta)

1498: \]

1499: generates $H_{n-1}(L_b)$.  

1500: \end{coro}

1501:    

1502: \pr Let $\t{\Delta}$ be a distinguished set of Lefschetz vanishing cycles as in 

1503: Theorem  ~\ref{16apr}. We can see easily that 

1504: $\tau(\t{\Delta})=\Delta\cup h(\Delta)\cup\cdots\cup h^{p-1}(\Delta)$.

1505: $\diamond$

1506: \\

1507: 

1508: The fiber $L_b$ does not contain the points of $\{F=0\}\cap\{G=0\}$, so this corollary partially claims that the cycle around a point of $\{F=0\}\cap\{G=0\}$ is a rational sum of vanishing cycles. 

1509: In the initial steps of this article my objective was to prove the following 

1510: corollary.

1511: \begin{coro} 

1512: Suppose that $n=2$ and the generic fiber of $\integ$ has genus greater than zero.  Then the action of the monodromy group on a vanishing cycle generates $H_{n-1}(L_b)$.

1513: Let $\omega_1$ be a meromorphic 1-form in the projective space of dimension two whose pole divisor is a union of some fibers of $\integ$. If

1514: \[

1515:   \int _{\delta_t} \omega _1 =0

1516: \]

1517: for a continuous family $\delta_t$ of vanishing cycles, then $\omega_1$ restricted to the closure of each fiber of $\integ$ is exact.

1518: \end{coro}

1519:  We recall that in the above corollary we have assumed  the generic conditions of Proposition ~\ref{gen}.

1520: \\

1521: \pr 

1522:  The first part is a direct consequence of Theorem ~\ref{monodromy} and Proposition ~\ref{a}. For the second part it is enough to prove that 

1523: \[

1524:   \int _{\delta_t} \omega _1 =0

1525: \]

1526: For all 1-cycles in the fibers of $\integ$.$\square$  

1527: \\

1528: Using the ramification map $\tau$, the reader can verify that:

1529: \begin{cla}

1530: \label{khaste}

1531: Let $D_0$ be a small disk around $0$ and $l$ be the straight line which

1532: connects 0 to $b_0$, a point in $\partial D_0$, then

1533: \begin{enumerate}

1534: \item

1535:  $(L_l,L_{b_0})$ is a strong deformation retract  of $(L_{D_0},L_{b_0})$;

1536: \item

1537: There is a $C^\infty$ function $\phi :l\times L_{b}\rightarrow L_{l}$ such that

1538:  $\phi$ is a fiber bundle trivialization on $l\backslash \{0\}$ and the

1539:   restriction of $\phi$ to $\{0\}\times L_{b_0}$, namely $g$, is a finite 

1540:   covering map  of degree $p$ from $L_{b_0}$ to $L_0$ ;

1541: \item

1542: There is a monodromy $h:L_{b_0}\rightarrow L_{b_0}$ around $0$ 

1543: such that for every 

1544: $x\in L_{b_0}$ we have 

1545: \[

1546: g^{-1}(g(x))=\{x,h(x),\cdots ,h^{p-1}(x)\}

1547: \]

1548: in particular $h^p=I$ and $g\circ h=g$.   

1549: \end{enumerate}

1550: \end{cla}

1551:   

1552: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%    

1553:  \appendix

1554: \section{Generic Properties}

1555: \label{secgeneric}

1556: \ \ \ \ \

1557: Here we will prove  Proposition ~\ref{gen}. The main tool is the 

1558: transversality theorem which appears both in Algebraic Geometry and Differential 

1559: Topology. We will work in the category of algebraic varieties but

1560: the whole of this

1561:  discussion can be done in the $C^{\infty}$ category of  manifolds.

1562:  \\

1563:   In the sequel by $TX$ we denote the tangent bundle of the variety $X$ 

1564:   and by $(TX)_0$ we denote the image of the zero section of the vector bundle 

1565:   $TX$. For any $x\in X$ we have 

1566: \[

1567: T_{0_x}(TX)=T_{0_x}(TX)_0\oplus T_{0_x}(T_xX)

1568: \]

1569: and so we can define

1570: \[

1571:  d:T_{0_x}(TX)\rightarrow  T_{0_x}(T_xX)

1572: \]

1573: $d$ is the projection on the second coordinate.

1574:  We will essentially use the following transversality theorem in algebraic

1575:   geometry:

1576: \begin{Theo}

1577: \label{transversality}

1578: Let $f:X\rightarrow Z$ and $\pi : X\rightarrow A$ be morphisms 

1579: ($C^\infty$ functions) between smooth varieties (resp. $C^\infty$ manifolds) 

1580: and $W$ be a smooth subvariety (resp. submanifold) of $Z$. 

1581: Also assume that $\pi$ is surjective and $f$ is transverse to $W$, then there 

1582: exists an open dense subset $U$ of $A$ such that 

1583: $f\mid _{\pi ^{-1}(\alpha )}$ is transverse to $W$ for 

1584: every $\alpha \in U$.  

1585: \end{Theo}

1586: %\pr $f$ is transverse to $W$ therefore $\t{W}=f^{-1}{W}$ is an smooth subvariety of $X$. Applying second Bertini theorem (Sard theorem) to $\t{\pi}=\pi\mid_{\t{W}}:\t{W}\rightarrow A$, we conclude that there is an open dense set $U$ in $A$ such that for all $\alpha\in U$, $\t{\pi}^{-1}(\alpha)$ is an smooth subvariety of $\t{W}$.   

1587: 

1588: \pr This theorem is a consequence of Bertini's theorems ( see \cite{sha}). For

1589: more information about the transversality theorem the

1590: reader is referred to \cite{spe} and \cite{abr}.  

1591: \\

1592: Recall that $f:X\rightarrow Z$ is transverse to $W$ if for every $x\in X$ 

1593: with $y=f(x)\in W$, we have $T_{y}W +(T_xf)(T_x(X))=T_y(Z)$. 

1594: This is equivalent to this fact that $f^{-1}(W)$ is empty or is a smooth

1595:  subvariety of $X$ of dimension $dim(X)-dim(Z)+dim(W)$. The following well-known proposition will be used.

1596: \begin{cla}

1597: \label{nondegenerate}

1598: Let $f: X\rightarrow Z $ be a morphism between two smooth varieties and $dim(Z)=1$. Then the critical points of $f$ are nondegenerate, if and only if, $Tf:TX\backslash{(TX)}_0 \rightarrow TZ$ is transverse to ${(TZ)}_0$.

1599: \end{cla}

1600: %\pr

1601: %In a local coordinate around $x$ and $y=f(x)$ we have:

1602: %\[

1603: %Tf:TX\rightarrow TZ

1604: %\]

1605: %\[

1606: %Tf(x,v)=(f(x),Df(x)(v))

1607: %\]

1608: %\[

1609: %T^2f:T^2X \rightarrow T^2Z

1610: %\]

1611: %\[

1612: %T^2f(x,v,u,w)=(f(x),Df(x)(v),Df(x)(u),D^2f(x)(v)(u)+Df(x)(w) )

1613: %\]

1614: %Let $(x,v)\in TM\backslash (TM)_0$ and 

1615: %$Tf(x,v)\in (TZ)_0$ i.e., $Df(x)(v)=0$ and  $v\not =0$ then $Tf$ is not 

1616: %transverse at $(x,v)$ if and only if 

1617: %\[ 

1618: %\forall u,w\in \C ^n ,\ D^2f(x)(v)(u)+Df(x)(w)=0

1619: %\]

1620: %which is equivalent to the fact that $Df(x)=0$ and $D^2f(x)(v)=0$, or in 

1621: %other words $x$ is a degenerate critical point of $f$. $\diamond$  

1622: 

1623:  Let 

1624: \[

1625: X=\{(F,G,x)\in \PP \times \pro \mid F(x)\not =0 ,\ G(x)\not =0\}

1626: \]

1627: \[

1628: g:X \rightarrow \S,\ g(F,G,x)=\integ (x)=f(x)

1629: \]

1630: and $\tilde{T} X$ be the subvector bundle of $TX$ whose fiber 

1631: $\tilde{T}_{(F,G,x)}X$ is the tangent space of $\{(F,G)\}\times \pro$.

1632:  Let also $\tilde{T}g$ be the restriction of $Tg$ to $\tilde{T} X$ and 

1633:  $\pi :\tilde{T}X\rightarrow \PP$ be the projection on the parameter $(F,G)$.

1634: \begin{cla}

1635: \label{22agu}

1636: For a generic pair $(F,G)$, the critical points of $\integ$ in

1637: $\pro\backslash(\{F=0\}\cup\{G=0\})$ are nondegenerate.

1638: \end{cla}

1639: 

1640: \pr

1641: According to the transversality theorem and Proposition 

1642: ~\ref{nondegenerate} it is enough

1643:  to prove that 

1644:  \[

1645:  \tilde{T}g: \tilde{T}X \rightarrow T\S

1646:  \]

1647: is transverse to $(T\S )_0$.  In a local coordinate around 

1648: $(F,G,x',v)\in \tilde{T}X$ we have

1649: \[

1650: \tilde{T}g(F,G,x',v)=(\integ (x'),D(\integ) (x')(v))

1651: \]

1652: \[

1653: B=d\circ T_{(F,G,x',v)}(\tilde{T}g)(\bar{F},\bar{G},u,w)=

1654: \]

1655: \[ D^2f(x')(v)(u)+Df(x')(w)+pD(f\frac{\bar{F}}{F})(x')(v)-qD(f\frac{\bar{G}}{G})(x')(v)

1656: \]

1657: If $\tilde{T}g$ is not transverse to $(T\S )_0$ at $(x' ,v)$ with 

1658: $v\not = 0$,  then $B=0$ for all $u,w,\bar{F},\bar{G}$. 

1659: Putting $\bar{F}=\bar{G}=0$ we get

1660: \[

1661: Df(x')=0,\ D^2f(x')(v)=0

1662: \]

1663: Let $x'=(x_1',x_2',\ldots x_n')$ and $v'=(v_1,v_2,\ldots v_n)$  

1664: then for all $i=1,2,\ldots ,n$ putting $\bar{F}=x-x'_i,\ \bar{G}=0$,  

1665: we obtain $v_i=0$. This implies that $v=0$ which is a contradiction.$\diamond$

1666: \\

1667: 

1668: The next step is to prove that generically the images of the critical points 

1669: of $\integ$ are distinct in $\S$. I did not succeed 

1670: to get this generic property by using the transversality theorem, 

1671: therefore I will prove it in the projective space of dimension two, by an elementary arguments in algebraic geometry. The proof in higher dimensions is the same. 

1672: The following lemmas will be used:

1673: \begin{lem}

1674:   \label{linear}

1675: Let $\phi :\C ^n\rightarrow \C ^m $ be a linear map and $A$ be a subvariety of

1676:  $\C ^m$. 

1677:  Then $A\cap Im(\phi)=A_1\cup A_2\cup\cdots$ is the decomposition of 

1678:  $A\cap Im(\phi)$ into irreducible components, if and only if, 

1679:  $\phi ^{-1}(A)=\phi ^{-1}(A_1)\cup \phi ^{-1}(A_2)\cup \cdots$ is the 

1680:  decomposition of  $\phi ^{-1}(A)$ into irreducible components.

1681: \end{lem}

1682: 

1683: \pr This is due to the fact that we can choose a basis for the vector space 

1684: $\C ^n$ such that

1685:  $\phi :\C ^n =\C ^{n-m'}\times \C ^{m'}\rightarrow Im(\phi)=\C ^{m'}$ 

1686:  is the projection on the second coordinate.$\diamond$

1687: \\

1688: 

1689: Let 

1690: \[

1691: A'=\{x\in \C ^6 \mid px_4x_2-qx_1x_5=px_4x_3-qx_1x_6=0\}

1692: \]

1693: \[

1694: A'_r=\{x\in \C ^6\mid x_4=x_1=0\}

1695: \]

1696: \[

1697: A'_c=\{x\in \C ^6 \mid px_4x_2-qx_1x_5=px_4x_3-qx_1x_6=x_2 x_6-x_3x_5=0\}

1698: \]

1699: \begin{lem}

1700: \label{ire}

1701: The following statements are true:

1702: \begin{enumerate}

1703: \item

1704: $A'$ has two irreducible components $A'_r$ and $A'_c$;

1705: \item

1706:  $A'\times A'$ has four irreducible components 

1707:  $A'_{ij}=A'_i\times A'_j,\ i,j=r,c$; 

1708: \item 

1709:  For any linear subspace $V$ of $\C^{12}$ of dimension greater than 8,

1710:   $A'\times A'\cap V$ has also four irreducible components 

1711:   $A'_{ij}\cap V,\ i,j=r,c$. 

1712: \end{enumerate}

1713: \end{lem}

1714: In fact, from this lemma we only need to the fact that, for any linear 

1715: subspace $V$ of $\C^{12}$ of dimension greater than 8, 

1716: $A'_{cc}\cap V$ is irreducible. 

1717: \\

1718: \def\pro2{{\Bbb C}P(2)}

1719: Consider an affine open set $\C ^2\subset \pro2$ and let 

1720: \[

1721: 0=(0,0),\ 1=(0,1)

1722: \]

1723: Define

1724: \[

1725: A=\{\omega=\INT\mid (F,G)\in \PP\ \&\ \omega   

1726: \hbox{ has singularity at 0 and 1}\}

1727: \]

1728: 

1729: \begin{lem}

1730: The variety $A$ has exactly four irreducible components 

1731: $A_{rr},\ A_{rc},A_{cr},A_{cc}$. The component $A_{rc}$ contains 

1732: all 1-forms in $A$ which have a radial singularity at $0$ and a center 

1733: singularity at $1$. In the same way other components are defined. 

1734: \end{lem}

1735: \pr

1736: For any $p\in \C ^2$ define the linear map

1737: \[

1738: 	\phi _{p}:\PP\rightarrow \C ^6

1739: 	,\ \phi _p(F,G)=(F(p),F_x(p),F_y(p),G(p),G_x(p),G_y(p))

1740: \]

1741: where the partial derivatives are considered in the fixed affine coordinate.

1742: Also we define 

1743: \[	

1744: 	\phi :\PP\rightarrow \C ^{12},\ \phi =(\phi _{(0,0)},\phi _{(0,1)} )

1745: \]

1746: 

1747: We can assume that $deg(F)\geq 2$ and $deg(G)\geq 1$. With this hypotheses 

1748: the reader can check that

1749: $dim(Im(\phi))\geq 8$. Now our assertion is the direct consequence of  Lemmas 

1750: ~\ref{linear}, ~\ref{ire}.

1751: \\

1752: 

1753: {\bf Proof of Proposition ~\ref{gen}:} According to Proposition ~\ref{22agu}, it is enough to prove that generically the image of nondegenerate critical points are distinct. Let

1754: \[

1755: S=\{(F,G)\in A_{cc} \mid \integ (0)=\integ (1)\} 

1756: \]

1757: Let $(F,G)\in\PP$ and $\integ$ have $r$ nondegenerate critical points 

1758: $p_1,\cdots ,p_r$. There is an small perturbation $(\bar{F},\bar{G})$ of

1759:  $(F,G)$ such that $\integb $ has r distinct critical values.

1760:   Suppose that this is not true, then we can assume that $\integ$ has maximal 

1761:   number $r'$ of critical values in some neighborhood of $(F,G)$ and $r'<r$. 

1762:   There exist two critical points $p_1,p_2$ of $\integ$ such that 

1763:   $\integ (p_1)=\integ (p_2)$ and for any $(\bar{F},\bar{G})$ near $(F,G)$ 

1764:   with corresponding critical point $\bar{p_1},\bar{p_2}$ near $p_1$ and $p_2$,

1765:    respectively, we have

1766: \[

1767: \integb (\bar{p_1})=\integb(\bar{p_2})

1768: \]

1769: Let $L$ be the linear automorphism of $\pro2$ which sends $0$ and $1$ to

1770:  $\bar{p_1}$ and $\bar{p_2}$, respectively. 

1771:  In some neighborhood $U$ of $(F\circ L,G\circ L )$ in $\PP$ we have 

1772:  $A_{cc}\cap U\subset S\cap U$. Since $A_{cc}$ is an irreducible variety we

1773:  conclude that 

1774:  $A_{cc}\subset S$ which is contradiction because 

1775: \[

1776: (xy^{a-1}+y^a, \frac{q}{p}-a +x+ay)\in A_{cc}\backslash S 

1777: \]

1778: $\diamond$

1779: 

1780: 

1781: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1782: \begin{thebibliography}{99}

1783: \bibitem{abr}

1784: R. Abraham, J. Robbin, Transversal Mappings and flows, W. A. Benjamin, Inc. New

1785: York, 1967.

1786: \bibitem{arn}

1787: V.I. Arnold, S.M. Gusein-zade, A.N. Varchenko, Singularities of Differential 

1788: Maps, V. II, Birkh\"auser, 1988.  

1789: \bibitem{che}

1790: D. Ch\'eniot, Vanishing Cycles in a Pencil of Hyperplane Sections of a 

1791: Nonsingular Quasi-projective Variety, Proc. London Math. Soc. (3)72, 1996, 

1792: 515-544.

1793: \bibitem{cer}

1794: D. Cerveau, A. Lins Neto, Irreducible Components of the Space of Holomorphic 

1795: Foliation of Degree 2 in $\pn ,\ N\geq 0$, Annals of mathematics, 143(1996), 

1796: 577-612.

1797: \bibitem{del}

1798: P. Deligne, Equations Diff\'erentielles \`a Points Singuliers R\'eguliers,

1799: Lecture Notes in Mathematics, 163, Springer-Verlag, 1970.

1800: \bibitem{dim}

1801: A. Dimca, Singularities and Topology of Hypersurfaces, Univeritext,

1802:  Springer-Verlag, New York Inc.(1992).

1803: \bibitem{ehr}

1804: C. Ehresmann, Sur l'espaces fibr\'es diff\'erentiables, C.R. Acad. Sci. Paris 

1805: 224(1947), 1611-1612.

1806: \bibitem{esn}

1807: H. Esnault, E. Viehweg, Lectures on Vanishing Theorems, Birkh\"auser, DMV

1808: Seminar Band 20, 1992.

1809: \bibitem{ily}

1810: Yu. S. Ilyashenko, The Origin of Limit Cycles Under Perturbation of Equation

1811: $\frac{dw}{dz}=-\frac{R_z}{R_w}$, Where $R(z,w)$ Is a Polynomial, 

1812: Math. USSR , Sbornik, Vol. 7, (1969), No. 3.

1813: \bibitem{kem}

1814: G. R. Kempf, Algebraic Varities, London Mathematical Society, Lecture Note

1815: series 172, Cambridge Univercity Press.

1816: \bibitem{lam}

1817: K. Lamotke, The Topology of Complex Projective Varieties After S. Lefschetz, 

1818: Topology, Vol 20, 1981. 

1819: \bibitem{lef}

1820: S. Lefschetz, L'Analysis Situs et la G\'eom\'etrie Alg\'ebrique Paris, 

1821: Gauthier-Villars, 1924.

1822: \bibitem{alci}

1823: A. Lins Neto, B. A. Sc\'ardua, Folea\c{c}\~oes Alg\'ebricas Complexas,

1824: $21^\circ$ Col\'oquio Brasileiro de Matem\'atica. 

1825: \bibitem{mas}

1826: W.S. Massey, A Basic Course in Algebraic Topology, 1991, Springer-Verlag, 

1827: New York.

1828: \bibitem{mil}

1829: J. Milnor, Morse theory, Princeton Press, Princeton, NJ, 1963.

1830: \bibitem{pau}

1831: E. Paul, Cycles \'Evanescents D'une Fonction de Liouville de Type

1832: $f_1^{\lambda_1}\cdots f_p^{\lambda_p}$, Ann. Inst. Fourier, Grenoble, 45, 1

1833: (1995), 31-63.

1834: \bibitem{sha}

1835: I. R. Shafarevich, Basic Algebraic Geometry, V. I, Springer-Verlag, Berlin Heidelberg,

1836: 1977,1994.

1837: \bibitem{spe}

1838: R. Speiser, Transversality Theorems For Families of Maps, Lecture Notes in

1839: Mathematics, 1311, Algebraic  Geometry Sundance, 1986.

1840: \bibitem{ste}

1841: N. Steenrod, The homology of Fiber Bundles, Princeton University press (1951).

1842: \end{thebibliography}

1843: \smallskip

1844: \rightline{Hossein Movasati}

1845: \rightline{Instituto de Matem\'atica Pura e Aplicada, IMPA}

1846: \rightline{Estrada Dona Castorina, 110, 22460-320}

1847: \rightline{Rio de Janeiro, RJ, Brazil}

1848: \rightline{ E-mail: hossein@impa.br}

1849: \end{document} 

1850: