m.gonzalez/Steenbrink
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

actualizar Mult3.tex

Browse Source
This commit is contained in:
m.gonzalez 2024-07-11 13:06:35 +02:00
parent 7b03d02b4f
commit 532d4c80b5
2 changed files with 71 additions and 38 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

BIN
Mult3.pdf
View File

Binary file not shown.

109
Mult3.tex
View File

@ -310,13 +310,13 @@ Consider a the germ of a multiplicity 3 singular point of hypersurface. An equat
\medskip
\noindent \emph{First step:} By means of a Tschirnhaus transformation, reduce $F$ to a monic Weierstrass equation $F = z^3 -3 b_2z + 2 b_3$ that does not have terms of degree 2 with respect to $z$. Assume that the coefficients of degree one and zero with respect to $z$ are both non zero, i.e., the product $b_2b_3 \ne 0$. If either $b_2 =0$ and $b_3 \ne 0$ or $b_2 \ne 0$ and $b_3=0$ it reduces to the ase of a suspension that we have studied elsewhere. See \cite{}.
\noindent \emph{First step:} By means of a Tschirnhaus transformation, reduce $F$ to a monic Weierstrass equation $F = z^3 -3 b_2z + 2 b_3$ that does not have terms of degree 2 with respect to $z$. Assume that the coefficients of degree one and zero with respect to $z$ are both non zero, i.e., the product $b_2b_3 \ne 0$. If either $b_2 =0$ and $b_3 \ne 0$ or $b_2 \ne 0$ and $b_3=0$ it reduces to the case of a suspension that we have studied elsewhere. See \cite{}.
%We propose the following general method to study germs of multiplicity 3 singular points of hypersurfaces. Firstly, consider an equation of the germ given by a monic Weierstrass polynomial $F = z^3 + a_1z^2 + a_2z + a_3 \in \CC[[{\bf x}]][z]$ of degree 3 with respect to $z$. Assume that $D_z F$ is not zero. We will be mainly interested in the case when $F$ is degenerate with respect to its Newton polyhedron. Otherwise, the partial resolution associated with the dual of the Newton polyhedron of $F$ gives plenty information of the geometry and topology of the singular point of the hypersurface defined by $F$. By means of a Tschirnhaus transformation, reduce $F$ to a monic Weierstrass equation $z^3 + b_2z + b_3$ that does not have terms of degree 2 with respect to $z$. Assume that the coefficients of degree one and zero with respect to $z$ are both non zero, i.e., the product $b_2b_3 \ne 0$.
\medskip
\noindent \emph{Second step:} Consider an embedded resolution $\pi: X \rightarrow \CC^2$ of the reduced divisor given by the product $G=b_2b_3D_zF$. Consider the usual stratification of the pair $(X, (\pi \circ G)^{-1} \{G=0\})$ defined in terms of decomposition $\cup_{j \in J} E_j$ of $(\pi \circ G)^{-1} \{G=0\}$ into irreducible components. Pulling back $F$ under the map $\pi \times id$, we obtain for each stratum $E_I^\circ$ a \emph{quasi-trinomial local expression} for $F$. Section \ref{3nomial} clasifies the 5 posible quasi-trinomial local forms having terms of degrees 3, 1, and 0 with respect to $z$. These are the quasi-trinomials forms that appear at this step because we have previously made a Tschirnhaus transformation. Notice that the local expression is quasi-trinomial due to the fact that $\pi$ resolves both $b_2$ and $b_3$. Moreover, the quasi-trinomial local form is quasi-ordinary because $\pi$ also resolves the discriminant $D_zF$. The local quasi-trinomial expression will be non degenerate with respect to its Newton polyhedron unless the monomials $M_1$, $M_2$ and their coefficients $c_1$ and $c_2$ satisfy a precise condition. In principle, this precise condition can be defined on a subset of a stratum of $E_I^\circ$, with $\emptyset \ne I \subset J$. If it is the case, we consider the corresponding refinement of the stratification. Moreover, the degeneracies of the degenerate trinomial local case are either on horizontal faces or of the form $(z-af(x,y))(z-bf(x,y))$ with $a, b \in \CC^\ast$, and $f \in \CC[x,y]$.
\noindent \emph{Second step:} Consider an embedded resolution $\pi: X \rightarrow \CC^2$ of the reduced divisor given by the product $G=b_2b_3D_zF$. Consider the usual stratification of the pair $(X, (\pi \circ G)^{-1} \{G=0\})$ defined in terms of decomposition $\cup_{j \in J} E_j$ of $(\pi \circ G)^{-1} \{G=0\}$ into irreducible components. Pulling back $F$ under the map $\pi \times id$, we obtain for each stratum $E_I^\circ$ a \emph{quasi-trinomial local expression} for $F$. Section \ref{3nomial} clasifies the 5 posible quasi-trinomial local forms having terms of degrees 3, 1, and 0 with respect to $z$. These are the quasi-trinomials forms that appear at this step because we have previously made a Tschirnhaus transformation. Notice that the local expression is quasi-trinomial due to the fact that $\pi$ resolves both $b_2$ and $b_3$. Moreover, the quasi-trinomial local form is quasi-ordinary because $\pi$ also resolves the discriminant $D_zF$. The local quasi-trinomial expression will be non degenerate with respect to its Newton polyhedron unless the monomials $M_1$, $M_2$ and their coefficients $c_1$ and $c_2$ satisfy a precise condition. {\color{red}In principle, this precise condition can be defined on a subset of a stratum of $E_I^\circ$, with $\emptyset \ne I \subset J$. If it is the case, we consider the corresponding refinement of the stratification.} Moreover, the degeneracies of the degenerate trinomial local case are either on horizontal faces or of the form $(z-af(x,y))(z-bf(x,y))$ with $a, b \in \CC^\ast$, and $f \in \CC[x,y]$.
\medskip
@ -402,6 +402,8 @@ The monic Weierstrass polynomial $z^3 -3b_2z + 2b_3$ is non degenerate unless $M
\section{Jung´s method for power series of multiplicity three }
\subsection{Embbeded resolution of the discriminant and classification of their exceptional divisors}
Consider an embedded resolution $\pi: X \rightarrow \CC^2$ of the reduced divisor given by the product $G=b_2b_3D_zF$. Consider the usual stratification of the pair $(X, (\pi \circ G)^{-1} \{G=0\})$ defined in terms of decomposition $\cup_{j \in J} E_j$ of $(\pi \circ G)^{-1} \{G=0\}$ into irreducible components, i.e., for each subset $I \subset J$ we define
$$E_I^\circ = \bigcap_{i \in I} E_j \setminus \bigcup_{j \not \in I} E_j .$$
@ -421,12 +423,17 @@ There are four possibilities for the behaviour of the numbers $N^0_j, N^\infty_j
\item $3 N^0_j = 2N^\infty_j < N^\Delta_j$ and in this case it is mapped constantly to $1$ under $\mathcal{M}$.
\end{itemize}
In order to distinguish between dicritical divisors and divisors that are mapped constantly to a value different than $0$, $1$ or $\infty$ one notices that connected components of dicritical divisors must intersect both to zeros and poles and connected components of divisors mapped constantly to a value different than $0$, $1$ or $\infty$ do not intersect zeros or poles but do intersect dicritical divisors.
\begin{remark}
In the third case, $N^\Delta_j$ is divisible by 6. In order to distinguish between dicritical divisors and divisors that are mapped constantly to a value different than $0$, $1$ or $\infty$, one notices that connected components of dicritical divisors must intersect both to zeros and poles and connected components of divisors mapped constantly to a value different than $0$, $1$ or $\infty$ do not intersect zeros or poles but do intersect dicritical divisors.
If the strata $E^\circ_{i,j} \ne \emptyset$ and $\mathcal{M}(E_i)=1$ then either $\mathcal{M}(E_j)=1$ or $E_j$ is dicritical, i.e., $\mathcal{M}(E_j)=\mathbb{P}^1$.
\end{remark}
\begin{definition} For a divisor of the last type, the difference
\begin{definition}\label{extriplepole} For a divisor of th last type, i.e., a divisor that is mapped constantly to $1$, the difference
$N^\Delta_j - 3 N^0_j = N^\Delta_j - 2 N^\infty_j$ is called the \emph{excess of the discriminant} and it is denoted by $M^\Delta_j$.
\end{definition}
@ -435,7 +442,7 @@ $N^\Delta_j - 3 N^0_j = N^\Delta_j - 2 N^\infty_j$ is called the \emph{excess of
\begin{example} Consider $F=z^3 - 3b_2z + 2b_3$ with
$b_2= ((y^2 + x^3)^2 + x^5y)((x^2 + y^3)^2 + y^5x)$
and
$b_3= (y^3+x^4)(x^3+y^4)$. The minimal resolution $\pi$ of the product $G=b_2 b_3 D_zF$ is obtained after 15 blow-ups. The following figure and table give the dual resolution graph of $G$ and its associated numerical data.
$b_3= (y^3+x^4)(x^3+y^4)$. The minimal resolution $\pi$ of the product $G=b_2 b_3 D_zF$ is obtained after 30 blow-ups. The following figure and table give the dual resolution graph of $G$ and its associated numerical data.
@ -453,7 +460,7 @@ $b_3= (y^3+x^4)(x^3+y^4)$. The minimal resolution $\pi$ of the product $G=b_2 b_
\coordinate (G5) at (-0.1,0.75);
\coordinate (G6) at (-0.2,0.75);
\coordinate (G7) at (0,0.75);
\node[above] at (G7) {{\small ${{F^\Delta_i}}_{{}_{i=1}}^{{}{13}}$}};
\node[above] at (G7) {{\small ${{F^\Delta_i}}_{{}_{i=16}}^{{}{28}}$}};
\coordinate (G8) at (0.1,0.75);
\coordinate (G9) at (0.2,0.75);
\coordinate (G10) at (0.3,0.75);
@ -485,17 +492,17 @@ $b_3= (y^3+x^4)(x^3+y^4)$. The minimal resolution $\pi$ of the product $G=b_2 b_
\coordinate (F1) at (1.2,0.75);
\node[above] at (F1) {{\small $F^{\infty}_1$}};
\node[above] at (F1) {{\small $F^{\infty}_{33}$}};
\coordinate (F2) at (1.95,0.75);
\node[above] at (F2) {{\small $F^\Delta_{14}$}};
\node[above] at (F2) {{\small $F^\Delta_{29}$}};
\coordinate (F3) at (6,0.75);
\node[above] at (F3) {{\small $F^{0}_1$}};
\node[above] at (F3) {{\small $F^{0}_{31}$}};
\coordinate (F4) at (-1.2,0.75);
\node[above] at (F4) {{\small $F^{\infty}_2$}};
\node[above] at (F4) {{\small $F^{\infty}_{34}$}};
\coordinate (F5) at (-1.95,0.75);
\node[above] at (F5) {{\small $F^\Delta_{15}$}};
\node[above] at (F5) {{\small $F^\Delta_{30}$}};
\coordinate (F6) at (-6,0.75);
\node[above] at (F6) {{\small $F^{0}_2$}};
\node[above] at (F6) {{\small $F^{0}_{32}$}};
@ -578,42 +585,45 @@ N^\Delta_i& 25& {\bf 26}& {\bf 52}& {\bf 78}& {\bf 104}& {\bf 130}& {\bf 156}& {
\hline
\hline
\nu_i&2&3&5&7&9&11&13&6&11\\
\chi()&-13&1&-1&2&2&2&-4&1&-1\\
\chi()&-11&1&-1&0&0&0&-2&1&-1\\
E_i^2& -7& -2& -2& -2& -2& -2& -1&-2&-1\\
\hline
\end{array}$$
The divisor $E_1$ is mapped to the value 1, the divisors $E_7$ and $E_{13}$ are dicriticals, and the rest of exceptionals divisors are poles. The 13 arrows $F^\Delta_i$ with $i=1, \dots, 13$ attached to the divisor $E_1$ correspond to irreducible components of the strict transforms of $D_zF$, the arrows attachef to the divisor $E_7$ (resp. $E_{13}$) correspond to an irreducible component strict transform of $b_3$ and another irreducible component of $D_zF$.
The divisor $E_1$ is mapped to the value 1, the divisors $E_7$ and $E_{15}$ are dicriticals, and the rest of exceptionals divisors are zeros. The 13 arrows $F^\Delta_i$ with $i=16, \dots, 29$ attached to the divisor $E_1$ correspond to irreducible components of the strict transforms of $D_zF$, the arrows attached to the divisor $E_7$ (resp. $E_{13}$) correspond to an irreducible component strict transform of $b_3$ and another irreducible component of $D_zF$. Finally the arrow attched to $E_9$ (resp. to $E_{17}$) correspond to the strict transform of $b_2$.
%$$\pi^{-1} \{\mathcal{M} = 0\} = 2 (F^{0}_1 + F^{0}_2) + 4 (E_2 + E_5 + E_{10} + E_{13}) + 6 (E_4 + E_{12} ) + 8 (E_3 + E_{11}) + 11 (E_8+E_{16}) + 22 (E_9 + E_{17})) , \quad \pi^{-1} \{\mathcal{M} = 1\} = \sum_{i=1}^{15} F^\Delta_i + E_1$$
$$\pi^{-1} \{\mathcal{M} = 0\} = 2 ( \sum_{1,2}F^{0}_i + \sum_{6,14} E_i ) + 4 \sum_{2,5,10,13}E_i + 6 \sum_{4,12 }E_i + 8 \sum_{3,11}E_i + 11\sum_{8,16} E_i + 22 \sum_{9,17} E_i ,$$
$$\pi^{-1} \{\mathcal{M} = \infty\} = 3 F^{\infty}_1 + 3 F^{\infty}_2, \quad \pi(E_7) = \pi(E_{15}) = \mathbb{P}^1, \quad \pi^{-1} \{\mathcal{M} = 1\} = \sum_{i=1}^{15} F^\Delta_i + E_1$$
$$\pi^{-1} \{\mathcal{M} = 0\} = 2 ( \sum_{31,32}F^{0}_i + \sum_{6,14} E_i ) + 4 \sum_{2,5,10,13}E_i + 6 \sum_{4,12 }E_i + 8 \sum_{3,11}E_i + 11\sum_{8,16} E_i + 22 \sum_{9,17} E_i ,$$
$$\pi^{-1} \{\mathcal{M} = \infty\} = 3 F^{\infty}_{33} + 3 F^{\infty}_{34}, \quad \pi(E_7) = \pi(E_{15}) = \mathbb{P}^1, \quad \pi^{-1} \{\mathcal{M} = 1\} = \sum_{i=16}^{30} F^\Delta_i + E_1$$
The stratification associated with the resolution $\pi$ consist of $35$ strata $E^\circ_{i,j}$ of dimension 0 and 8 strata $E_i^\circ$ of dimension 1 and $\chi(E_i^\circ) \ne 0$.
The former correspond to the 16 pairs $(i,j)$ listed belllow
$$\begin{array}{cccccccc}
(1,7)& (2,3)& (3,4)& (3,9)& (4,5)& (5,6)& (6,7)& (8, 9) \\
(1,15)& (10,11)& (11,12)& (11,17)& (12, 13)& (13,14)& (14,15)& (16,17)
\end{array}$$
the 13 intersections $E_1 \cap F_i^\Delta$ for $i=1, \dots, 13$ and the 6 intersections
$$ E_7 \cap F_1^\infty, \quad E_7 \cap F_{14}^\Delta, \quad E_9 \cap F_1^0, \quad E_{15} \cap F_2^\infty, \quad E_{15} \cap F_{14}^\Delta, \quad E_{17} \cap F_2^0 $$
The stratification associated with the resolution $\pi$ consist of $35$ strata $E^\circ_{i,j}$ of dimension 0 and 11 strata $E_i^\circ$ of dimension 1 and $\chi(E_i^\circ) \ne 0$.
The former correspond to the elements $(i,j)$ of the set
$$S= \left \{ \begin{array}{ccccccccc}
(1,7)& (1,15)& (1,16)& (1,17)& (1,18)& (1,19)& (1,20)& (1,21)& (1,22)\\
(1,23)& (1,24)& (1, 25)& (1, 26)& (1, 27)& (1, 28)& (2,3)& (3,4)& (3,9)\\
(4,5)& (5,6)& (6,7)& (7, 29)& (7, 33)& (8, 9)& (9,31)& (10,11)& (11,12)\\
(11,17)& (12, 13)& (13,14)& (14,15)& (15,30)& (15, 34)& 16,17)& (17,32)
\end{array} \right \} $$
%the 13 intersections $E_1 \cap F_i^\Delta$ for $i=16, \dots, 28$ and the 6 intersections
%$$ E_7 \cap F_{34}^\infty, \quad E_7 \cap F_{30}^\Delta, \quad E_9 \cap F_{32}^0, \quad E_{15} \cap F_{35}^\infty, \quad E_{15} \cap F_{31}^\Delta, \quad E_{17} \cap F_{33}^0 $$
and the latter correspond with the indexes
$$1, \quad 2, \quad 3, \quad 7, \quad 8, \quad 10, \quad 11, \quad 15.$$
$$1, \quad 2, \quad 3, \quad 7, \quad 8, \quad 9, \quad 10, \quad 11, \quad 15, \quad 16, \quad 19.$$
Notice that the minimal resolution $\pi'$ of the discriminant $D_zF$ is obtained after just 11 blow-ups. The blowups corresponding to divisors $E_8, E_9, E_{14}$ and $E_{15}$ are needed to resolve $b_2$.
\end{example}
Let us consider the birational map $\tilde{\pi} := ( \pi, id ): \tilde{X} : =X \times \CC \rightarrow \CC^{3}$, where $id: \CC \rightarrow \CC$ is the identity map, and the pullback of $F$ under $\pi \times id$. The pullback $\pi^\ast F$ can be expressed in terms of the pullbacks of $b_2$ and $b_3$ as
\subsection{Trinomial forms associated to the strata}
Let us consider the birational map $\tilde{\pi} := ( \pi, id ): \tilde{X} : =X \times \CC \rightarrow \CC^{3}$, where $id: \CC \rightarrow \CC$ is the identity map, and the pullback of $F$ under $\pi \times id$. The pullback $\tilde{\pi}^\ast F$ can be expressed in terms of the pullbacks of $b_2$ and $b_3$ as
\begin{equation}
z^3 -3 \pi^\ast b_2 z + 2 \pi^\ast b_3.
\end{equation}
Over each non empty strata $E^\circ_I$ of $X$ created by the resolution $\pi$,
for the pullback of $F$ under the birational map $\pi \times id$.
Since $\pi$ is a resolution of both $b_2$ and $b_3$, over each non empty strata $E^\circ_I$ of $X$ created by the resolution $\pi$, the terms $(\pi^\ast b_3)_{|_I}$ and $(\pi^\ast b_3)_{|_I}$ can be reexpressed in terms of the multiplicities $N^f_i$ and $N^g_i$ with $i \in I$ as
over each non empty strata $E^\circ_I$ of $X$ created by the resolution $\pi$.
Since $\pi$ is a resolution of both $b_2$ and $b_3$, over each non empty strata $E^\circ_I$ of $X$ created by the resolution $\pi$, the terms $(\pi^\ast b_3)_{|_I}$ and $(\pi^\ast b_3)_{|_I}$ can be reexpressed in terms of the multiplicities $N^0_i$ and $N^\infty_i$ with $i \in I$ as
\begin{equation}
(\pi^\ast b_2)_{|_I}= x_1^{N^0_{i_1}} y_1^{N^0_{i_2}} u_1(x_1, y_1) \quad {\rm and } \quad
(\pi^\ast b_3)_{|_I}= x_1^{N^\infty_{i_1}} y_1^{N^\infty_{i_2}} u_1(x_1, y_1)
@ -652,7 +662,7 @@ in the case $I= \{i\}$
and
$$\begin{cases}
z^3 + x^{N^\infty_i}y^{N^\infty_j}, & \mathcal{M}(E_i), \mathcal{M}(E_j) \ne 1, \infty \\
z^3 + x^{\frac{N^0_i}{2}}y^{\frac{N^0_j}{2}}z^2 + x^{N_i^\infty + M_i^\Delta}y^{N_j^\infty + M_ij^\Delta} , & \mathcal{M}(E_i) = \mathcal{M}(E_j) = 1 \\
z^3 + x^{\frac{N^0_i}{2}}y^{\frac{N^0_j}{2}}z^2 + x^{N_i^\infty + M_i^\Delta}y^{N_j^\infty + M_j^\Delta} , & \mathcal{M}(E_i) = \mathcal{M}(E_j) = 1 \\
z^3 + x^{\frac{N^0_i}{2}}y^{\frac{N^0_j}{2}} z^2 + x^{N_i^\infty + M_i^\Delta}y^{N_j^\infty} , & \mathcal{M}(E_i) = 1, \mathcal{M}(E_j) = \mathbb{P}^1 \\
z^3 + x^{N^0_{i}}y^{N^0_{j}} z + x^{N^\infty_{i}}y^{N^\infty_{j}}, & {\rm else}\\
\end{cases}$$
@ -988,9 +998,9 @@ Therefore, $F$ is non degenerate with respect to its Newton polyhedron if and on
\section{Contibutions to the topological zeta function and the A´Campo monodromy zeta function of the trinomial forms}
\begin{proposition}
The monodromy zeta function and the local topological zeta function $Z(F,s)_0$ of $F$ equal
$$(1 - t^2)^{\alpha} \prod_{I} \zeta_{f_I}(t)^{-\chi(E_I^\circ)}$$
\begin{proposition}\label{formestrpi}
The monodromy zeta function $\zeta_F(t)$ and the local topological zeta function $Z(F,s)_0$ of $F$ equal
$$(1 - t^3)^{\alpha} \prod_{I} \zeta_{f_I}(t)^{-\chi(E_I^\circ)}$$
$$\sum_I \chi(E^\circ_I) Z(f_I, \omega_I)_{o_I}$$
where $I$ varies over the statification associated to the resolution $\pi$ of $b_2b_3D_zF$, the exponent $\alpha$ is the number of points of blow-ups needed to obtain $\pi$, and $o_I$ is a point in $E^\circ_I$.
@ -998,9 +1008,12 @@ $$\sum_I \chi(E^\circ_I) Z(f_I, \omega_I)_{o_I}$$
\end{proposition}
\begin{example}
\end{example}
\begin{example} The resolution $\pi$ of Example \ref{extriplepole} consist of $\alpha=15$ blow-ups and the associated stratification has 35 strata. According to Proposition \ref{formestrpi} we have
$$\zeta_F(t) = \frac{(1-t^3)^{15} \zeta^{}_{f_{2}} \zeta^{2}_{f_{4}} \zeta^{2}_{f_{5}} \zeta^{2}_{f_{6}} \zeta^{}_{f_{8}} \zeta^{}_{f_{10}} \zeta^{2}_{f_{12}} \zeta^{2}_{f_{13}} \zeta^{2}_{f_{14}} \zeta^{}_{f_{16}} }
{\zeta^{13}_{f_{1}} \zeta^{}_{f_{3}} \zeta^{4}_{f_{7}} \zeta^{}_{f_{9}} \zeta^{}_{f_{11}} \zeta^{4}_{f_{15}} \zeta^{}_{f_{17}} }$$
$$Z(F,s)_0=-11Z(f_1,s)_0 + \sum_{i=2,8,10,16}Z(f_i,s)_0 - \sum_{i=3, 9,11,17} Z(f_i,s)_0 - 2 \sum_{i=7,15}Z(f_i,s)_0 + \sum_{(i,j) \in S} Z(f_{i,j},s)_0,$$
where $S$ is the subset indexing the 0-dimensional strata of Example \ref{extriplepole} and we have omitted the symbols of differential forms associated to each strata for shortness sake.
\end{example}
@ -1208,7 +1221,27 @@ $$\frac{1}{\nu_1\nu_2(3s+1)} - \frac{2}{(3m_2s + 2\nu_1+m_2)(3n_2s + 2\nu_2+m_2)
\subsection{} Third case $f= z^3 - x^{m_1}y^{n_1}z^2 - x^{m_3}y^{n_3}$ and such that $3m_1 \leq m_3, 3 n_1 \leq n_3$ and $3(m_1, n_1) \ne 2(m_3, n_3)$.
\section{Reductions}
\subsection{First reduction: Elimination of the 0-dimensional strata due to the strict transforms of $b_2$ and of $b_3$} The following Proposition shows that the o-dimensional strata associated to the strict transforms of $b_2$ and $b_3$ can eliminated from the computations of \ref{}. Each of these points can be added to the 1-dimensional strata satisfying that its clousere contains it, and the invariant is conputed with the trinomial form of the 1-dimensional strata.
\begin{proposition}\label{red1}Assume that $E_i$ is an exceptional divisor of the resolution $\pi$ with $\pi(E_i)=0$ (resp. $\pi(E_i)=\infty$) and that $F^0_j$ is an strict transform of $b_2$ (resp. $b_3$) such that $E_i \cap F_j \ne \emptyset$. Then
$$Z(F_i, \omega_i)_{o_i}=Z(F_{i,j}, \omega_{i,j})_{o_{i,j}} $$
\end{proposition}
\begin{example} The proposition \ref{red1} can be applied to the strata
$$E^\circ_{9,32}=E_9 \cap F^0_{32}, \quad E^\circ_{17,33}=E_{17} \cap F^0_{33}, \quad E^\circ_{7,34}= E_7 \cap F^\infty_{34} \quad \hbox{ and } \quad E^\circ_{15,35}=E_{15} \cap F^\infty_{35}$$
of Example \ref{extriplepole}. For instance, ifrom the formulas given in the previous section; we have
$$Z(F_9, \omega_9)_{o_9}=Z(F_{9,31}, \omega_{9,31})_{o_{9,31}}$$
$$ Z(F_{15}, \omega_{15})_{o_{15}}=Z(F_{15,34}, \omega_{15,34})_{o_{15,34}}$$
\end{example}
\subsection{Second reduction: Elimination of connected components of exceptional divisors containing a $(-1)$-curve}
After performing the First reduction, connected components of exceptional divisors
\begin{proposition}\label{red2}
\end{proposition}
\section{On generic projections}