2000 Mathematics Subject Classification. 14H45.
<ph f="cmbx">Gieseker-Petri special curves.</ph>

### Abel Castorena.

Abstract.- In the moduli space of smooth and irreducible projective curves of genus $g$  , ${\mathcal{ℳ}}_{g}$  , let $\mathcal{G}{\mathcal{P}}_{g}$  be the locus of curves that do not satisfy the Gieseker-Petri theorem and let $\mathcal{G}{\mathcal{P}}_{g,d}^{1}$  be the subvariety of $\mathcal{G}{\mathcal{P}}_{g}$  formed by curves $C$  of genus $g$  with a $|D|={g}_{d}^{1}$  free of base points for which ${\mu }_{D}:{H}^{0}\left(C,D\right)\otimes {H}^{0}\left(C,{K}_{C}\left(-D\right)\right)\to {H}^{0}\left(C,{K}_{C}\right)$  is not injective. In this work we show that $\mathcal{G}{\mathcal{P}}_{g,g-2}^{1}$  and $\mathcal{G}{\mathcal{P}}_{g,g-3}^{1}$  are irreducible divisors in ${\mathcal{ℳ}}_{g}$  . As consequence of this results we show that $\mathcal{G}{\mathcal{P}}_{g}$  is a divisor for $g=8,9,10$  .
0.Introduction Let ${\mathcal{ℳ}}_{g}$  be the moduli space of smooth and irreducible projective curves of genus $g$  .
Let $C\in {\mathcal{ℳ}}_{g}$  and let ${K}_{C}$  be the canonical bundle of $C$  . The Gieseker-Petri theorem (see [8], p.285) says that for every line bundle $L$  on a general curve $C\in {\mathcal{ℳ}}_{g}$  , the Petri map ${\mu }_{L}:{H}^{0}\left(C,L\right)\otimes {H}^{0}\left({K}_{C}\otimes {L}^{-1}\right)\to {H}^{0}\left(C,{K}_{C}\right)$  is injective. This implies that the locus $\mathcal{G}{\mathcal{P}}_{g}:=\left\{C\in {\mathcal{ℳ}}_{g}|C\text{does not satisfy the Gieseker-Petri theorem}\right\}$  is a proper closed Zariski subset in ${\mathcal{ℳ}}_{g}$  . A curve that do not satisfy he Gieseker-Petri theorem is called Gieseker-Petri special curve. Let $C\in {\mathcal{ℳ}}_{g}$  and $L\to C$  a line bundle of degree $d$  with $r+1={h}^{0}\left(C,L\right)$  . The Brill-Noether number is defined as $\rho \left(g,r,d\right):={h}^{0}\left(C,{K}_{C}\right)-{h}^{0}\left(C,L\right){h}^{0}\left(C,{K}_{C}\otimes {L}^{-1}\right)=g-\left(r+1\right)\left(g-d+r\right)$  . When $\rho \left(g,d,r\right)<0$  , the Petri map ${\mu }_{L}$  is not injective, and in this case Steffen showed that each component of ${\mathcal{ℳ}}_{g,d}^{r}:=\left\{C\in {\mathcal{ℳ}}_{g}|{G}_{d}^{r}\left(C\right)\ne \varnothing \right\}$  has codimension at most $-\rho \left(g,r,d\right)$  in ${\mathcal{ℳ}}_{g}$  (see [9]). When $\rho =-1$  , in [6] Eisenbud and Harris showed that ${\mathcal{ℳ}}_{g,d}^{r}$  has a unique irreducible component of codimension one in ${\mathcal{ℳ}}_{g}$  . M. Teixidor showed (see [10],[11]) that the locus ${\mathcal{ℳ}}_{g}^{1}:=\left\{C\in {\mathcal{ℳ}}_{g}|C\text{has a autoresidual}{g}_{g-1}^{1}\right\}$  is an irreducible divisor in ${\mathcal{ℳ}}_{g}$  . For integers $g,r,d\ge 1$  such that $\rho \ge 0$  consider the variety ${G}_{d}^{r}\left(C\right):=\left\{\left(L,V\right):V\subset {H}^{0}\left(C,L\right),\text{dim}V=r+1\right\}$  . Denote ${\mu }_{V}:V\otimes {H}^{0}\left(C,K\otimes {L}^{-1}\right)\to {H}^{0}\left(C,K\right)$  . We define $\mathcal{G}{\mathcal{P}}_{g,d}^{r}:=\left\{C\in {\mathcal{ℳ}}_{g}:\text{there exists a base point free}\left(L,V\right)\in {G}_{d}^{r}\left(C\right)\text{with}{\mu }_{V}\text{not injective}\right\}\subset \mathcal{G}{\mathcal{P}}_{g}$  .
Eisenbud and Harris showed that when $\rho \left(g,r,d\right)=0$  , $\mathcal{G}{\mathcal{P}}_{g,d}^{r}$  is a divisor in ${\mathcal{ℳ}}_{g}$  (see [5], section 5). Following this notation, G. Farkas showed (see [7]) that for integers $g\ge 4$  and $\frac{g+2}{2}\le k\le g-1$  , $\mathcal{G}{\mathcal{P}}_{g,k}^{1}$  has a divisorial component. The existence of such component is obtained in an inductive step in the following sense: assuming that $\mathcal{G}{\mathcal{P}}_{g-1,k-1}^{1}$  has a divisorial component $Z$  , for which a general curve $C\in Z$  has a isolated base point free $|D|={g}_{k-1}^{1}$  with ${\mu }_{D}$  not injective. Consider now $p\in C$  a general point and an elliptic curve $E$  , and let ${X}_{0}=:C{\cup }_{p}E$  be a compact type curve. The author shows that the curves ${X}_{0}$  fill up a divisor on $\overline{{\Delta }_{1}}$  and that ${X}_{0}\in \overline{\mathcal{G}{\mathcal{P}}_{g,k}^{1}}\cap \overline{{\Delta }_{1}}$  , so $\mathcal{G}{\mathcal{P}}_{g,k}^{1}$  has a divisorial component ${Z}^{\prime }$  . This component is abstract in the sense that the author describes only elements on ${\overline{\mathcal{G}\mathcal{P}}}_{g,k}^{1}-\mathcal{G}{\mathcal{P}}_{g,k}^{1}$  .
Starting with the case $k\ge 3,g=2k-2$  , the locus $\mathcal{G}{\mathcal{P}}_{2k-2,k}^{1}$  is a divisor in ${\mathcal{ℳ}}_{g}$  by ([5], section 5). So applying the above inductive step the author contructs divisorial components $\mathcal{G}{\mathcal{P}}_{2k-2+a,k+a}^{1}$  for all $k\ge 3$  and $a\ge 0$  .
In general is not easy describe all irreducible components of $\mathcal{G}{\mathcal{P}}_{g}$  , instead for specific low genus we can describe the irreducible components of $\mathcal{G}{\mathcal{P}}_{g}$  , for example in [3], the author use an degeneration argument to show that $\mathcal{G}{\mathcal{P}}_{7}$  is a divisor with components ${\mathcal{ℳ}}_{7}^{1},{\mathcal{ℳ}}_{7,4}^{1},\mathcal{G}{\mathcal{P}}_{7,5}^{1}$  . In this case $\mathcal{G}{\mathcal{P}}_{7,5}^{1}$  is a irreducible divisor. This divisorial component is the image of the subvariety of the Severy variety formed by curves of geometric genus seven and degree seven with eight double points, seven of them lying on a conic. In this work we generalize this contruction in the following way giving an explicit description of the elementes of $\mathcal{G}{\mathcal{P}}_{g,g-2}^{1}$  :
Let $C$  be a non tetragonal curve $C$  of genus $g$  with a primitive $|D|={g}_{g-2}^{1}$  such that the residual ${g}_{g}^{2}$  of the ${g}_{g-2}^{1}$  induces a birational map onto a plane curve $\Gamma$  of degree $g$  and geometric genus $g$  with $\delta =\frac{\left(g-1\right)\left(g-2\right)}{2}-g$  double points as singularities.
In proposition 2.1 we show that ${\mu }_{D}$  is not injective if and only if there exists a curve $G$  of degree $g-5$  containing $\delta -1$  double points. Consider the Severi variety ${\mathcal{V}}^{g,g}$  of plane curves of degree $g$  and geometric genus $g$  having only double points as singularities. We consider the subvariety ${\mathcal{V}}_{\delta }^{g,g}\subset {\mathcal{V}}^{g,g}/PGL\left(3,\mathbb{C}\right)$  formed by plane curves with exactly $\delta =\frac{\left(g-1\right)\left(g-2\right)}{2}-g=\frac{g\left(g-5\right)}{2}+1$  double points. Let ${\mathcal{V}}_{0}:=\left\{\Gamma \in {\mathcal{V}}_{\delta }^{g,g}:\delta -1\text{double points lie on a curve of degree}g-5\right\}$  . In lemma 2.4 we show that for a non tetragonal curve $C$  of genus $g$  there is at most finitely many line bundles free of base points $L\in {W}_{g-2}^{1}\left(C\right)$  for which the Petri map ${\mu }_{L}$  is not injective.
As a consequence of proposition 2.1 and lemma 2.4 we obtain that the image of the natural morphism ${\mathcal{V}}_{0}\to {\mathcal{ℳ}}_{g}$  is an irreducible codimension one component in ${\mathcal{ℳ}}_{g}$  , that is, if ${\mathcal{V}}_{0}\ne \varnothing$  , we have that $\mathcal{G}{\mathcal{P}}_{g,g-2}^{1}$  is an irreducible divisor in ${\mathcal{ℳ}}_{g}$  . We enclose this as one of the main theorems of this work:
Theorem 1.- $\mathcal{G}{\mathcal{P}}_{g,g-2}^{1}$  is an irreducible divisor in ${\mathcal{ℳ}}_{g}$  .
For $g\le 10$  , ([4]) assures that $\mathcal{G}{\mathcal{P}}_{g,g-2}^{1}\ne \varnothing$  .
To show that $\mathcal{G}{\mathcal{P}}_{g,g-3}^{1}$  is a irreducible divisor, in lemma 2.5 we show that if $C$  is non tetragonal and $\mathcal{ℰ}\in {W}_{g-2}^{1}\left(C\right)$  is such that has only a point $p$  as base locus and ${\mu }_{\mathcal{ℰ}}$  not injective, then $\mathcal{ℒ}:=\mathcal{ℰ}+p$  is a theta characteristic on $C$  . In corollary 2.6 we obtain that there is at most finitely many $\mathcal{ℰ}\in {W}_{g-2}^{1}\left(C\right)$  as in the hypothesis of lemma 2.5 with ${\mu }_{\mathcal{ℰ}}$  is not injective. So we obtain in proposition 2.8 that there is at most finitely many $L\in {W}_{g-3}^{1}$  free of base points for which ${\mu }_{L}$  is not injective.
Thus we have our second main theorem:
Theorem 2.- $\mathcal{G}{\mathcal{P}}_{g,g-3}^{1}$  is a irreducible divisor in ${\mathcal{ℳ}}_{g}$  .
As an application of our main theorems we show in section 3 that $\mathcal{G}{\mathcal{P}}_{g}$  is a divisor for $g=8,9,10$  .
1.Preliminaries 1.1. Let $C$  be a smooth and irreducible curve of genus $g$  , $D$  a divisor on $C$  and $K$  the canonical bundle on $C$  . We say that the linear series $|D|$  is primitive if $|D|$  and $|K-D|$  are non empty and free of base points.
Lemma 1.1. If there exists $|D|$  such that ${\mu }_{D}:{H}^{0}\left(C,D\right)\otimes {H}^{0}\left(C,K-D\right)\to {H}^{0}\left(C,K\right)$  is not injective then there exists a primitive series $|{D}^{\prime }|$  such that ${\mu }_{{D}^{\prime }}$  is not injective.
Proof: see ([3]).
We will make use of The Base point free pencil trick ([2], p.126):
Let $L$  be an invertible sheaf on an algebraic curve $C$  . Let $\mathcal{ℱ}$  a torsion free ${\mathcal{O}}_{C}-$  module on $C$  . Let ${s}_{1},{s}_{2}$  sections lineary independents of ${H}^{0}\left(C,L\right)$  . Let $\mathcal{ℬ}$  the base locus of the space $V$  generated by ${s}_{1},{s}_{2}$  . Then the kernel of the multiplication map $V\otimes {H}^{0}\left(C,\mathcal{ℱ}\right)\to {H}^{0}\left(C,\mathcal{ℱ}\otimes L\right)$  is isomorphic to ${H}^{0}\left(C,\mathcal{ℱ}\otimes {L}^{-1}\left(\mathcal{ℬ}\right)\right)$  .
1.2. Let $C$  be a smooth curve of genus g and $\mathcal{U}\subset {\mathcal{ℳ}}_{g}$  small open neighbourhood of the point representing $C$  in ${\mathcal{ℳ}}_{g}$  . There exist a finite cover $\stackrel{~}{\mathcal{U}}$  of $\mathcal{U}$  and a scheme ${\mathcal{W}}_{d}^{r}$  parametrizing pairs $\left({C}^{\prime },{L}^{\prime }\right)$  , ${C}^{\prime }$  parametrized by $\stackrel{~}{\mathcal{U}}$  and ${L}^{\prime }\in {\text{Pic}}^{d}\left({C}^{\prime }\right)$  with ${h}^{0}\left({C}^{\prime },{L}^{\prime }\right)\ge r+1$  . Let $\stackrel{~}{\mathcal{G}{\mathcal{P}}_{g,d}^{r}}:=\left\{\left(C,L\right)\in {\mathcal{W}}_{d}^{r}:\text{rank}{\mu }_{L}\le g-\left(\rho +1\right)\right\}$  . We have the commutative diagram: $\stackrel{~}{\mathcal{G}{\mathcal{P}}_{g,d}^{r}}-----\to \text{inclusion}{\mathcal{W}}_{d}^{r}\pi ↓\pi ↓\mathcal{G}{\mathcal{P}}_{g,d}^{r}-----\to \text{inclusion}{\mathcal{ℳ}}_{g}$  Let $\gamma :=\text{codimension of}\stackrel{~}{\mathcal{G}{\mathcal{P}}_{g,d}^{r}}$  , then $\gamma \le \rho +1$  . We have three basic questions:
1.Does each irreducible component of $\stackrel{~}{\mathcal{G}{\mathcal{P}}_{g,d}^{r}}$  of codimension $\gamma$  map in ${\mathcal{ℳ}}_{g}$  with general fibre of dimension $\rho +1-\gamma$  ?
2.Is $\gamma =\rho +1$  and $\text{dim}\stackrel{~}{\mathcal{G}{\mathcal{P}}_{g,d}^{r}}=3g-3+\left(\rho -\gamma \right)=3g-4$  ?
3.Is $\pi :\stackrel{~}{\mathcal{G}{\mathcal{P}}_{g,d}^{r}}\to \mathcal{G}{\mathcal{P}}_{g,d}^{r}$  generically finite?
In the next section we will to construct components ${\stackrel{~}{\mathcal{G}\mathcal{P}}}_{g,d}^{1}$  for $d=g-2,g-3$  that gives afirmative answer to the above questions.
2.The locus $\mathcal{G}{\mathcal{P}}_{g,g-2}^{1}$  and $\mathcal{G}{\mathcal{P}}_{g,g-3}^{1}$  .
Let $C\in {\mathcal{ℳ}}_{g}$  be a non-tetragonal. We study primitive linear pencils on $C$  of degree $d$  for $d=g-2,g-3$  for which the Petri map is not injective. Asumme that the residual ${g}_{g}^{2}$  of the primitive linear series ${g}_{g-2}^{1}$  on $C$  induces a birational map onto a plane curve in ${\mathbb{P}}^{2}$  . Let $\Gamma$  be such a curve and $f:C\to \Gamma$  the normalization of $\Gamma$  .
We denote by ${△}_{\Gamma }$  the scheme of singular points of $\Gamma$  and $△:={f}^{*}\left({△}_{\Gamma }\right)$  , note that $△$  is a divisor of degree $2\delta$  . By the genus formula the lenght of $\left({△}_{\Gamma }\right)=\delta$  , i.e. ${△}_{\Gamma }$  is a curvilinear scheme consisting of $\delta$  double points which can be infinitely near.
We only consider the case where all singularities of $\Gamma$  are distincts.
Proposition 2.1.
(a). Let $\Gamma$  be a non tetragonal plane curve of degree $g$  and geometric genus $g$  with only double points as singularities and $f:C\to \Gamma$  its normalization. Suppose that there is curve $G$  of degree $g-5$  such that the scheme theoretic intersection of $G$  with ${△}_{\Gamma }$  has lenght equal to $\delta -1$  , i.e. ${f}^{*}\left(G\right)$  contains a divisor of degree $2\delta -2$  contained in $△$  , then $C$  does not satisfy the Gieseker-Petri theorem.
(b). Conversely if $C$  is non-tetragonal curve of genus $g$  such that there is a $|D|={g}_{g-2}^{1}$  on $C$  for which ${\mu }_{D}$  is not injective, then there is in ${\mathbb{P}}^{2}$  a birational model $\Gamma$  of $C$  of degree $g$  with only double points as singularities such that the ${g}_{g-2}^{1}$  is cut out by lines passing through a double point $p$  and there is a curve $G$  of degree $g-5$  such that $G$  contains ${△}_{\Gamma }-\left\{p\right\}$  .
Proof: First we will show (a). I will consider the most complicated case in which the support of ${△}_{\Gamma }=\left\{x\right\}$  . The other cases are easier and can be left to the reader.
If the support of ${△}_{\Gamma }=\left\{x\right\}$  , then $\Gamma$  has $\delta$  infinitely near double points. Let $\eta :={f}^{*}\left(x\right)$  . $\eta$  is a divisor of degree two and $△=\delta \eta$  . Our hypothesis means that the pullback ${f}^{*}G$  on $C$  contains $\left(\delta -1\right)\eta$  . Consider the $|D|={g}_{g-2}^{1}$  cut out on $C$  by the lines through $x$  . Let ${\ell }_{1},{\ell }_{2}$  be general such lines, cutting out on $C$  two effective divisors ${D}_{1},{D}_{2}\in |D|$  . The pullback of $G+{\ell }_{1}+{\ell }_{2}$  contains $\left(\delta +1\right)\eta +{D}_{1}+{D}_{2}\sim \left(\delta +1\right)\eta +2D$  . By adjunction formula([2],p. 53), ${K}_{C}\sim {\mathcal{O}}_{C}\left(g-3\right)\left(-△\right)$  , we have that ${K}_{C}\left(-2D\right)$  is effective, and since $\text{ker}{\mu }_{D}\simeq {H}^{0}\left(C,{K}_{C}\left(-2D\right)\right)$  , we have the assertion.
We show (b) as follow: Consider ${g}_{g}^{2}=|{K}_{C}\left(-D\right)|$  , where $|D|={g}_{g-2}^{1}$  . This linear system determines a birational morphism $C\to \Gamma \subset {\mathbb{P}}^{2}$  . By assumption $\Gamma$  has only double points. Since $C$  fails the Gieseker-Petri theorem for the ${g}_{g-2}^{1}$  , we have that $\text{ker}{\mu }_{D}\simeq {H}^{0}\left(C,{K}_{C}\left(-2D\right)\right)$  , but $|{K}_{C}\left(-2D\right)|\sim {g}_{g}^{2}-{g}_{g-2}^{1}$  is effective, so neccesary the ${g}_{g-2}^{1}$  is cut out by a pencil of lines through a singular point $p$  of $\Gamma$  . $\square$  2.2. In general it is complicated to construct an irreducible and reduced plane curve of degree $g$  and geometric genus $g$  with projective model as in proposition 2.1(a). However at least for $g\le 10$  such kind of curves exists. In ([4], p.148-156), the author show the existence of canonical surfaces in ${\mathbb{P}}^{3}$  with ${p}_{g}=4$  , degree $d=6,7,8,9,10$  and sectional genus $g=7,8,9,10,11$  with ordinary singularities.
The general plane section is semicanonical with number of nodes $\delta =3,7,12,18,25$  lying respectively on a curve of degree $1,2,3,4,5$  . A tangent general section has degree $d=6,7,8,9,10$  and the corresponding genus is $g=6,...,10$  . Such curves have respectively nodes $\delta =4,8,13,19,26$  where $3,7,12,18,25$  lying respectively on a curve of degree $1,2,3,4,5$  .
2.3. Let $C$  be a non tetragonal curve of genus $g$  . By Mumford theorem ([1], p. 193) the dimension of ${W}_{g-2}^{1}\left(C\right)$  is exactly $\rho \left(g,g-2,1\right)=g-6$  . Consider the morphism ${\text{Pic}}^{d}\left(C\right)\to {\text{Pic}}^{2d}\left(C\right)$  given by $L\to {L}^{2}$  inside the Jacobian $J\left(C\right)$  of $C$  . Note that this morphism has finite kernel. This implies that the subvariety ${X}_{1}:=\left\{{L}^{2}:L\in {W}_{g-2}^{1}\left(C\right)\right\}$  has dimension $\rho =g-6$  and ${T}_{L}\left({W}_{g-2}^{1}\left(C\right)\right)\simeq {T}_{{L}^{2}}{X}_{1}$  inside ${H}^{1}\left(C,{\mathcal{O}}_{C}\right)$  . Let $L\in {W}_{g-2}^{1}\left(C\right)$  be such that ${\mu }_{L}$  is not injective. Since $C$  is not hyperelliptic ${h}^{0}\left(C,{K}_{C}\otimes {L}^{-2}\right)=1$  , so there exist points $p,q\in C$  such that ${K}_{C}\otimes {L}^{-2}={\mathcal{O}}_{C}\left(p+q\right)$  with ${h}^{0}\left(C,{\mathcal{O}}_{C}\left(p+q\right)\right)=1$  , then ${h}^{0}\left(C,{K}_{C}\left(-p-q\right)\right)=g-2$  . If $p=q$  , then $L+p$  is a theta characteristic. So we only consider the case $p\ne q$  . Using the pairing $<,>$  of the Serre duality we have the description of the tangent space to the determinantal varieties ${W}_{d}^{r}$  ([1], p.186-187): ${T}_{L}\left({W}_{g-2}^{1}\left(C\right)\right)=\left(\text{Im}\left({\mu }_{L}\right){\right)}^{\perp }$  , where $\left(\text{Im}\left({\mu }_{L}\right){\right)}^{\perp }=\left\{\phi \in {H}^{1}\left(C,{\mathcal{O}}_{C}\right):<\phi ,{\mu }_{L}\left(s\otimes t\right)>=0,\text{for every}s\in {H}^{0}\left(C,L\right)\text{and every}t\in {H}^{0}\left(C,{K}_{C}\otimes {L}^{-1}\right)\right\}$  . By Serre duality we have that $<\phi ,{\mu }_{L}\left(s\otimes t\right)>=<\phi \cdot s,t>$  for every $s\in {H}^{0}\left(C,L\right)$  and every $t\in {H}^{0}\left(,{K}_{C}\otimes {L}^{-1}\right)$  . Define inside the Jacobian of $C$  , $J\left(C\right)$  , the subvariety ${X}_{2}:={K}_{C}-{W}_{2}\left(C\right)=\left\{{K}_{C}-\left(p+q\right):p+q\in {W}_{2}\left(C\right)\right\}\subset {W}_{2g-4}^{r}\left(C\right)$  for $r={h}^{0}\left(C,{K}_{C}\left(-p-q\right)\right)-1=g-3$  . We have that the dimension of ${X}_{2}=2$  .
Let $\mathcal{ℒ}={K}_{C}-\left(p+q\right)\in {X}_{2}$  be any point, then ${K}_{C}-\mathcal{ℒ}=p+q$  . The image of $\left({\mu }_{\mathcal{ℒ}}:{H}^{0}\left(C,\mathcal{ℒ}\right)\otimes {H}^{0}\left(C,{K}_{C}\otimes {\mathcal{ℒ}}^{-1}\right)\to {H}^{0}\left(C,{K}_{C}\right)\right)$  is equal to ${H}^{0}\left(C,{K}_{C}\left(-p-q\right)\right)$  , since ${h}^{0}\left(C,{\mathcal{O}}_{C}\left(p+q\right)\right)=1$  , ${\mu }_{\mathcal{ℒ}}$  is injective and ${T}_{\mathcal{ℒ}}{X}_{2}$  is a two dimensional subspace of ${H}^{1}\left(C,{\mathcal{O}}_{C}\right)\equiv {T}_{0}\left(J\left(C\right)\right)$  .
Lemma 2.4. Let $C$  be a non-tetragonal curve of genus $g$  . Suposse that there exist a primitive linear series $|L|={g}_{g-2}^{1}$  such that the residual ${g}_{g}^{2}$  induces a birational morphism from $C$  to a plane curve $\Gamma$  of degree $g$  in ${\mathbb{P}}^{2}$  with ${x}_{1},...,{x}_{\delta }$  nodes all distinct and ${x}_{1},...,{x}_{\delta -1}$  lying on curve of degree $g-5$  . Then there are at most finitely many line bundles $L\in {W}_{g-2}^{1}\left(C\right)$  free of base points for which ${\mu }_{L}$  is not injective.
Proof. We will show that ${X}_{1}\cap {X}_{2}$  is a finite set, where ${X}_{1}$  and ${X}_{2}$  are the subvarieties of $J\left(C\right)$  defined in 2.3. Let $L$  be like in the hypothesis, that is, we can assume that $|L|={g}_{g-2}^{1}$  is cut out by lines through the node ${x}_{\delta }$  of $\Gamma$  with $\text{ker}{\mu }_{L}\ne 0$  .
We have that ${L}^{2}\in {X}_{1}\cap {X}_{2}$  . Note that showing that ${T}_{{L}^{2}}{X}_{1}\cap {T}_{{L}^{2}}{X}_{2}=\left\{0\right\}$  inside ${H}^{1}\left(C,{\mathcal{O}}_{C}\right)$  we obtain that $L\in {W}_{g-2}^{1}\left(C\right)$  is an isolated point and this implies that ${X}_{1}\cap {X}_{2}$  is finite.
Consider the normalization map $f:C\to \Gamma$  . Let ${f}^{*}\left({x}_{\delta }\right)=\left\{p,q\right\}$  be for some points $p,q\in C$  , where $p\ne q$  because ${x}_{\delta }$  is a node. Since ${L}^{2}\in {X}_{1}\cap {X}_{2}$  , we have that ${K}_{C}\otimes {L}^{-2}\simeq {\mathcal{O}}_{C}\left(p+q\right)$  and ${h}^{0}\left(C,{K}_{C}\otimes {L}^{-2}\right)=\text{dim ker}{\mu }_{L}=1$  . We have that $\text{dim}{T}_{{L}^{2}}{X}_{1}=\text{dim}{T}_{L}\left({W}_{g-2}^{1}\left(C\right)\right)=\rho +\text{dim ker}{\mu }_{L}=g-5$  and $\text{dim}{T}_{{L}^{2}}{X}_{2}=2$  , then ${T}_{{L}^{2}}{X}_{1}\cap {T}_{{L}^{2}}{X}_{2}=\left\{0\right\}$  if and only if $\left({T}_{{L}^{2}}{X}_{1}{\right)}^{\perp }+\left({T}_{{L}^{2}}{X}_{2}{\right)}^{\perp }$  generates all of ${H}^{0}\left(C,{K}_{C}\right)$  , where $\perp$  means orthogonal complement with respect to Serre duality pairing $<,>$  . But $\text{dim}\left({T}_{{L}^{2}}{X}_{1}{\right)}^{\perp }+\text{dim}\left({T}_{{L}^{2}}{X}_{2}{\right)}^{\perp }=g+3$  , that is, $\text{dim}\left({T}_{{L}^{2}}{X}_{1}{\right)}^{\perp }+\text{dim}\left({T}_{{L}^{2}}{X}_{2}{\right)}^{\perp }-{h}^{0}\left(C,{K}_{C}\right)=3$  . So $\left({T}_{{L}^{2}}{X}_{1}{\right)}^{\perp }+\left({T}_{{L}^{2}}{X}_{2}{\right)}^{\perp }$  generates all of ${H}^{0}\left(C,{K}_{C}\right)$  if $p$  and $q$  impose independent conditions to $\text{image}{\mu }_{L}\subset {H}^{0}\left(C,{K}_{C}\right)$  , that is, if $\text{dimension of}\mathfrak{L}\left(-p-q\right)=3$  , where $\mathfrak{L}\left(-p-q\right):=\text{image}{\mu }_{L}\cap {H}^{0}\left(C,{K}_{C}\left(-p-q\right)\right)$  . We denote by $|\text{image}{\mu }_{L}|$  the linear system determined by the vector space $\text{image}{\mu }_{L}\subset {H}^{0}\left(C,{K}_{C}\right)$  Claim. The dimension of $\mathfrak{L}\left(-p-q\right)=3$  .
Proof of the claim. Let $D\in |{K}_{C}-L|$  not containing $p+q$  , and consider $D+|L|:=\left\{D+E:E\in |L|\right\}\subseteq |\text{image}{\mu }_{L}|$  , then if $p+q$  impose independent conditions to $D+|L|$  , then $p+q$  impose independent conditions to the linear system $|\text{image}{\mu }_{L}|$  , and in this case $\text{the dimension of}\mathfrak{L}\left(-p-q\right)=3$  . Let $\ell$  be a line through the node ${x}_{\delta }$  determined by $p$  and $q$  . Then the intersection $\ell \cdot \Gamma$  , of $\ell$  with $\Gamma$  , is $\ell \cdot \Gamma =p+q+{E}_{\ell }$  , where ${E}_{\ell }\in |L|$  . Since ${x}_{\delta }$  is a node there exists two different lines ${\ell }_{p},{\ell }_{q}$  through the node such that ${\ell }_{p}=2p+q+{E}_{p}$  , where $q$  is not in the support of the divisor ${E}_{p}$  , and ${\ell }_{q}=2q+p+{E}_{q},$  with $p$  not in the support of ${E}_{q}$  . Then ${E}_{p}+p\in |L|$  is the unique divisor given by the tangent line ${\ell }_{p}$  to the branch through $p$  not containing $q$  . Simillary $q+{E}_{q}$  is the unique divisor given by the tangent line ${\ell }_{q}$  thorugh the branch $q$  not containing the point $p$  . So we have that $p+q$  impose independent conditions to $D+|L|$  . $\square$  Proof of Theorem 1. We will show that $\pi \left({\stackrel{~}{\mathcal{G}\mathcal{P}}}_{g,g-2}^{1}\right)=\mathcal{G}{\mathcal{P}}_{g,g-2}^{1}$  is an irreducible divisor in ${\mathcal{ℳ}}_{g}$  .
Let $C$  be a non-tetragonal curve of genus $g$  and suppose that there is a line bundle $L\in {W}_{g-2}^{1}\left(C\right)$  free of base points such that the ${g}_{g}^{2}=|{K}_{C}\otimes {L}^{-1}|$  induces a projective model as in proposition 2.1, in other words, $\left(C,L\right)\in {\stackrel{~}{\mathcal{G}\mathcal{P}}}_{g,g-2}^{1}\subset {\mathcal{W}}_{g-2}^{1}$  . Since $C$  is non tetragonal we have that $\text{ker}{\mu }_{L}$  is one dimensional, so ${\mu }_{L}$  has rank five. In a small open neighbourhood ${U}_{C}$  inside the locus of non tetragonal curves containing $C$  in ${\mathcal{ℳ}}_{g}$  , we have a finite cover $\stackrel{~}{U}$  and pairs $\left({C}^{\prime },{L}^{\prime }\right)\in {\mathcal{W}}_{g-2}^{1},{C}^{\prime }\in \stackrel{~}{U},{L}^{\prime }\in {\text{Pic}}^{g-2}\left({C}^{\prime }\right)$  such that locally the Petri map is a homomorphism of vector bundles $\mu {|}_{\left({C}^{\prime },{L}^{\prime }\right)}:{H}^{0}\left({C}^{\prime },{L}^{\prime }\right)\otimes {H}^{0}\left({C}^{\prime },{K}_{{C}^{\prime }}\otimes \left({L}^{\prime }{\right)}^{-1}\right)\to {H}^{0}\left({C}^{\prime },{K}_{{C}^{\prime }}\right)$  For each $\left({C}^{\prime },{L}^{\prime }\right)\in {\stackrel{~}{\mathcal{G}\mathcal{P}}}_{g,g-2}^{1}$  , the homomorphism $\mu {|}_{\left({C}^{\prime },{L}^{\prime }\right)}$  has rank five, and in this case the subvariety ${\stackrel{~}{\mathcal{G}\mathcal{P}}}_{g,g-2}^{1}$  has codimension $\le g-5=\rho +1$  , that is, $\text{dim}{\stackrel{~}{\mathcal{G}\mathcal{P}}}_{g,g-2}^{1}\ge 3g-3+\rho -\gamma =3g-4$  . The lemma 2.4 implies that the projection $\pi :{\stackrel{~}{\mathcal{G}\mathcal{P}}}_{g,g-2}^{1}\to {\mathcal{ℳ}}_{g}$  is generically finite, then we have that $\pi \left({\stackrel{~}{\mathcal{G}\mathcal{P}}}_{g,g-2}^{1}\right)=\mathcal{G}{\mathcal{P}}_{g,g-2}^{1}$  is an irreducible divisor in ${\mathcal{ℳ}}_{g}$  . $\square$  .
Lemma 2.5. Let $C$  be a non tetragonal curve of genus $g$  with a $L\in {W}_{g-2}^{1}\left(C\right)$  with base locus $\mathcal{ℬ}=\left(p\right)$  such that ${\mu }_{L}$  is not injective, then $\mathcal{ℒ}:=L+p$  is a theta characteristic on $C$  .
Proof. Since $C$  is non tetragonal, by the base point free pencil trick we have that $\text{dim ker}{\mu }_{L}={h}^{0}\left(C,\left({K}_{C}\otimes {L}^{-1}\right)\otimes {L}^{-1}\left(p\right)\right)=1$  , then there exists ${q}_{1},{q}_{2},{q}_{3}\in C$  such that ${K}_{C}\otimes {L}^{-2}\left(p\right)\sim {q}_{1}+{q}_{2}+{q}_{3}$  . By Riemann-Roch we have that ${h}^{0}\left(C,{K}_{C}\left(3p\right)\right)=g+2$  and ${h}^{0}\left(C,{K}_{C}\left(3p-{q}_{2}-{q}_{3}\right)\right)=g$  . Note that ${K}_{C}+3p-{q}_{2}-{q}_{3}\sim 2L+2p+{q}_{1}=2\mathcal{ℒ}+{q}_{1}$  , then ${h}^{0}\left(C,{\mathcal{ℒ}}^{2}+{q}_{1}\right)=g$  . Since ${q}_{1}$  is base point of ${\mathcal{ℒ}}^{2}+{q}_{1}$  we have that ${h}^{0}\left(C,\left({\mathcal{ℒ}}^{2}+{q}_{1}\right)\right)={h}^{0}\left(C,\left({\mathcal{ℒ}}^{2}+{q}_{1}\right)-{q}_{1}\right)={h}^{0}\left(C,{\mathcal{ℒ}}^{2}\right)=g$  . By Riemann-Roch we have that ${h}^{0}\left(C,{K}_{C}\otimes {\mathcal{ℒ}}^{-2}\right)=1$  , that is ${K}_{C}\simeq {\mathcal{ℒ}}^{2}$  , then $\mathcal{ℒ}=L+p$  is a theta characteristic on $C$  .
Corollary 2.6.. Let $C$  and $L$  be as in the hypothesis of lemma 2.5. Then there is at most finitely many such $L\in {W}_{g-2}^{1}\left(C\right)$  for which ${\mu }_{L}$  is not injective.
Proof. Let $p$  the base locus of $L$  . For each such $L$  we have by lemma 2.5 that $L+p$  is a theta characteristic. Let ${L}_{0}$  and ${L}_{1}$  as in lemma 2.5 with ${p}_{0}$  and ${p}_{1}$  the base locus of ${L}_{0}$  and ${L}_{1}$  respectively. Note that if $|{L}_{0}|\ne |{L}_{1}|$  then $|{L}_{0}+{p}_{0}|\ne |{L}_{1}+{p}_{1}|$  . Since on an algebraic curve there are finitely many theta characteristics the assertion follows. $\square$  .
2.7. Now consider a non tetragonal curve $C$  of genus $g$  with a primitive pencil $|L|={g}_{g-3}^{1}$  for which ${\mu }_{L}$  is not injective. By the base-point-free pencil trick we have that $\text{dim ker}{\mu }_{L}={h}^{0}\left(C,{K}_{C}\otimes {L}^{-2}\right)=1$  . This implies that $|{K}_{C}\otimes {L}^{-2}|\ne \varnothing$  , that is, there exists points ${p}_{1},{p}_{2},{p}_{3},{p}_{4}$  on $C$  such that ${K}_{C}\otimes {L}^{-2}\sim {\mathcal{O}}_{C}\left({p}_{1}+{p}_{2}+{p}_{3}+{p}_{4}\right)$  with ${h}^{0}\left(C,{\mathcal{O}}_{C}\left({p}_{1}+{p}_{2}+{p}_{3}+{p}_{4}\right)\right)=1$  . Note that if (for example) ${p}_{1}={p}_{2}$  , then ${K}_{C}\otimes {L}^{-2}\otimes {O}_{C}\left(2{p}_{1}\right)\sim {\mathcal{O}}_{C}\left({p}_{3}+{p}_{4}\right)$  , that is, $\mathcal{ℰ}:=L\otimes {\mathcal{O}}_{C}\left({p}_{1}\right)\in {W}_{g-2}^{1}\left(C\right)$  is such that ${K}_{C}\otimes {\mathcal{ℰ}}^{-2}\sim {\mathcal{O}}_{C}\left({p}_{3}+{p}_{4}\right)$  . This case was analized in proof of lemma 2.4.
Thus we may assume that ${p}_{1},{p}_{2},{p}_{3},{p}_{4}\in C$  are all differentes. Note also that if $\alpha$  is the non-zero section of ${H}^{0}\left(C,{\mathcal{O}}_{C}\left({p}_{2}+{p}_{3}\right)\right)$  and $\beta$  is the non-zero section of ${H}^{0}\left(C,{\mathcal{O}}_{C}\left({p}_{4}\right)\right)$  , then $\alpha \otimes \beta$  is the non-zero section of ${H}^{0}\left(C,{\mathcal{O}}_{C}\left({p}_{2}+{p}_{3}+{p}_{4}\right)\right)$  . Proposition 2.8. There is at most finitely many line bundles $L\in {W}_{g-3}^{1}\left(C\right)$  free of base points for which ${\mu }_{L}$  is not injective.
Proof. Let $L\in {W}_{g-3}^{1}\left(C\right)$  free of base points for which ${\mu }_{L}$  is not injective. Then $K\otimes {L}^{-2}\sim {\mathcal{O}}_{C}\left({p}_{1}+{p}_{2}+{p}_{3}+{p}_{4}\right)$  as in 2.7. Let $\mathcal{ℰ}:=L\otimes {\mathcal{O}}_{C}\left({p}_{1}\right)$  be, then $\mathcal{ℰ}\in {W}_{g-2}^{1}\left(C\right)$  and $\mathcal{ℬ}={p}_{1}$  is the base locus of $\mathcal{ℰ}$  . Denote $\mathcal{ℱ}={K}_{C}\otimes {\mathcal{ℰ}}^{-1}$  . By the base point free pencil trick we have that $\text{ker}{\mu }_{\mathcal{ℰ}}\simeq {H}^{0}\left(C,\mathcal{ℱ}\otimes {\mathcal{ℰ}}^{-1}\left(\mathcal{ℬ}\right)\right)={H}^{0}\left(C,\left({K}_{C}\otimes {\mathcal{ℰ}}^{-1}\right)\otimes {\mathcal{ℰ}}^{-1}\left({p}_{1}\right)\right)={H}^{0}\left(C,{K}_{C}\otimes {L}^{-2}\left(-2{p}_{1}\right)\otimes {\mathcal{O}}_{C}\left({p}_{1}\right)\right)={H}^{0}\left(C,{\mathcal{O}}_{C}\left({p}_{2}+{p}_{3}+{p}_{4}\right)\right)$  , then $\text{dim ker}{\mu }_{\mathcal{ℰ}}={h}^{0}\left(C,{\mathcal{O}}_{C}\left({p}_{2}+{p}_{3}+{p}_{4}\right)\right)=1$  . This show that if there is a infinitely many line bundles $L\in {W}_{g-3}^{1}\left(C\right)$  free of base points for which ${\mu }_{L}$  is not injective, then there exists infinitely many line bundles $\mathcal{ℰ}\in {W}_{g-2}^{1}\left(C\right)$  each one of them with exactly one point as base locus with ${\mu }_{\mathcal{ℰ}}$  not injective. By corollary 2.6 this is impossible, so we have the assertion.
The proof of our second main theorem is similar to the proof of theorem 1. Our second theorem says:
Theorem 2. $\mathcal{G}{\mathcal{P}}_{g,g-3}^{1}$  is a irreducible divisor in ${\mathcal{ℳ}}_{g}$  .
3.The Gieseker-Petri locus $\mathcal{G}{\mathcal{P}}_{g}$  for $g=8,9,10.$  In this section we will show that for $g=8,9,10$  , $\mathcal{G}{\mathcal{P}}_{g}$  is a divisor. For this we will use our theorems and the the following result due to Castelnuovo (see proof [2, p.116]):
Castelnuovo's bound. Let $C$  be a smooth curve that admits a birational mapping onto a non-degenerated curve of degree $d$  in ${\mathbb{P}}^{r}$  . Then the genus $g$  of $C$  satisfies the inequality $g\le \pi \left(d,r\right)$  , where the Castelnuovo number $\pi \left(d,r\right)$  is defined by $\pi \left(d,r\right)=\frac{m\left(m-1\right)}{2}\left(r-1\right)+m\epsilon$  , where $m=\left[\frac{d-1}{r-1}\right]$  with $d-1=m\left(r-1\right)+\epsilon$  . Here $\left[s\right]$  denotes the integer part of the number $s$  .
Consider the stratification of ${\mathcal{ℳ}}_{g}$  given by gonality: ${\mathcal{ℳ}}_{g,2}^{1}\subseteq {\mathcal{ℳ}}_{g,3}^{1}\subseteq \cdot \cdot \cdot \subseteq {\mathcal{ℳ}}_{g,d}^{1}\subseteq \cdots \subseteq {\mathcal{ℳ}}_{g}$  where ${\mathcal{ℳ}}_{g,d}^{1}={\mathcal{ℳ}}_{g}$  for $d\ge \left[\frac{g+3}{2}\right]$  . Thus the number $\left[\frac{g+3}{2}\right]$  is the generic gonality for curves of genus $g$  . Note that a curve $C$  of genus $g$  with a ${g}_{d}^{1}$  belongs to $\mathcal{G}{\mathcal{P}}_{g,d}^{1}$  if $g>2\left(d-1\right)$  .
3.1. First we analize $\mathcal{G}{\mathcal{P}}_{8}$  . We recall that it is enough to consider primitive linear series $|D|={g}_{d}^{r}$  for which the Petri map ${\mu }_{D}$  is not injective for $2r\le d\le g-1=7$  . 3.1.1. $r=1,d=2,3,4,5$  . By gonality we have that ${\mathcal{ℳ}}_{8,2}^{1}\subseteq {\mathcal{ℳ}}_{8,3}^{1}\subseteq {\mathcal{ℳ}}_{8,4}^{1}\subseteq {\mathcal{ℳ}}_{8,5}^{1}$  . In this case $\rho \left(8,1,4\right)=-2<\rho \left(8,1,5\right)=0$  . By ([9]) we have that ${\mathcal{ℳ}}_{8,4}^{1}$  is of codimension at most two. Also by ([5], p. 360-361) we have that $\mathcal{G}{\mathcal{P}}_{8,5}^{1}$  is a divisor.
3.1.2. $r=1,d=6$  . By the theorem 1, $\mathcal{G}{\mathcal{P}}_{8,6}^{1}$  is an irreducible divisorial component of $\mathcal{G}{\mathcal{P}}_{8}$  3.1.3. $r=1,d=7$  . The locus of curves with a autoresidual $|D|={g}_{7}^{1}$  is the Teixidor divisor ${\mathcal{ℳ}}_{8}^{1}$  .
So by 3.1.1, 3.1.2, 3.1.3 we have that $\mathcal{G}{\mathcal{P}}_{8,5}^{1},\mathcal{G}{\mathcal{P}}_{8,6}^{1},{\mathcal{ℳ}}_{8}^{1}$  are divisorial components of $\mathcal{G}{\mathcal{P}}_{8}$  .
3.1.4. $r=2,d=4,5,6,7$  . A curve $C$  of genus eight with a ${g}_{4}^{2}$  is hyperelliptic, and by genus formula, there is no a curve of genus eight with a ${g}_{5}^{2}$  .
Let $C$  a curve with a primitive linear series ${g}_{6}^{2}$  . If ${g}_{6}^{2}$  induces a birational map $\psi :C\to Y\subset {\mathbb{P}}^{2}$  , then $Y:=\psi \left(C\right)$  is of degree six. By genus formula, either $Y$  has two double points or one tacnode. In any case the lines through the double point cut out a ${g}_{4}^{1}$  , so $C\in {\mathcal{ℳ}}_{8,4}^{1}$  . If the ${g}_{6}^{2}$  is not birational, then either $Y$  has degree two or three, so $C$  is either bielliptic or trigonal, in any case $C\in {\mathcal{ℳ}}_{8,4}^{1}\subset \mathcal{G}{\mathcal{P}}_{8,5}^{1}$  .
Curves of genus eight with a ${g}_{7}^{2}$  form a divisor since $\rho \left(8,2,7\right)=-1$  (see [6]). So for this BrillNoether number we obtain the Eisenbud-Harris component ${\mathcal{ℳ}}_{8,7}^{2}$  .
3.1.5. $r=3,d=6,7$  . Let $C$  be a curve of genus eight either with a ${g}_{6}^{3}$  or ${g}_{7}^{3}$  , By Castelnuovo's bound, $C$  either has a ${g}_{6}^{2}$  or ${g}_{7}^{2}$  , so by 3.1.4, $C$  belongs either to $\mathcal{G}{\mathcal{P}}_{8,5}^{1}$  or ${\mathcal{ℳ}}_{8,7}^{2}$  .
Thus we have that $\mathcal{G}{\mathcal{P}}_{8}$  is a divisor in ${\mathcal{ℳ}}_{8}$  where their divisorial components are $\mathcal{G}{\mathcal{P}}_{8,5}^{1},\mathcal{G}{\mathcal{P}}_{8,6}^{1},{\mathcal{ℳ}}_{8}^{1},{\mathcal{ℳ}}_{8,7}^{2}$  .
3.2. For $\mathcal{G}{\mathcal{P}}_{9}$  we have to analize for which primitive linear series $|D|={g}_{d}^{r}$  , ${\mu }_{D}$  is not injective for the values $2r\le d\le g-1=8$  3.2.1. $r=1,d=2,3,4,5,6,7,8$  . By gonality, curves with a ${g}_{d}^{1}$  for $d=2,3,4$  are contained in ${\mathcal{ℳ}}_{9,5}^{1}$  . Since $\rho \left(9,1,5\right)=-1$  . In this case we obtain that ${\mathcal{ℳ}}_{9,5}^{1}$  is a irreducible component of codimension one of $\mathcal{G}{\mathcal{P}}_{9}$  .
By theorem 2 we have that for $r=1,d=6$  , the locus of curves of genus nine with a $|D|={g}_{6}^{1}$  for which ${\mu }_{D}$  is not injective is the locus $\mathcal{G}{\mathcal{P}}_{9,6}^{1}$  . By theorem 1, curves of genus nine with a $|D|={g}_{7}^{1}$  for which ${\mu }_{D}$  is not injective is the irreducible divisor $\mathcal{G}{\mathcal{P}}_{9,7}^{1}$  .
Curves of genus nine with a autoresidual $|D|={g}_{8}^{1}$  is the Teixidor locus ${\mathcal{ℳ}}_{9}^{1}$  .
So for pencils we obtain the components ${\mathcal{ℳ}}_{9,5}^{1},\mathcal{G}{\mathcal{P}}_{9,6}^{1},\mathcal{G}{\mathcal{P}}_{9,7}^{1},{\mathcal{ℳ}}_{8}^{1}$  .
3.2.2. $r=2,d=4,5,6,7,8$  . Curves of genus nine with a ${g}_{2}^{4}$  are hyperelliptic.
There is no curve of genus nine with a ${g}_{5}^{2}$  . In analogous way to the case of genus eight note that a curve $C$  of genus nine with ${g}_{6}^{2}$  belongs to ${\mathcal{ℳ}}_{9,4}^{1}\subset {\mathcal{ℳ}}_{9,5}^{1}$  .
The locus of curves $C$  with a ${g}_{7}^{2}$  has codimension $\le 3$  since $\rho \left(9,2,7\right)=-3$  , (see [9]). The plane model induced by the ${g}_{7}^{2}$  is a curve $\Gamma$  of degree seven with six singularities. The lines through a singular point $p$  of $\Gamma$  cut out a ${g}_{7-s}^{1}$  , where $s$  is the multiplicity of the singular point $p$  , so $C\in {\mathcal{ℳ}}_{9,7-s}^{1}\subseteq {\mathcal{ℳ}}_{9,5}^{1}$  .
The locus of curves with a $|D|={g}_{8}^{2}$  for which ${\mu }_{D}$  is not injective is the divisorial component $\mathcal{G}{\mathcal{P}}_{9,8}^{2}$  since $\rho \left(9,2,8\right)=0$  ([5], p. 360-361). Thus for linear series of dimension two we obtain the codimension one components ${\mathcal{ℳ}}_{9,5}^{1},\mathcal{G}{\mathcal{P}}_{9,8}^{2}$  .
3.2.3. $r=3,d=6,7,8$  . By Castelnuovo's bound for $g=9,r=3,d=6,7,$  , a curve $C$  of genus nine with a ${g}_{d}^{3}$  is contained in a plane $H\simeq {\mathbb{P}}^{2}\subset {\mathbb{P}}^{3}$  , for $d=6,7$  . So, for such values $g,d,r$  we have that $C\in {\mathcal{ℳ}}_{9,5}^{1}$  by 3.2.2.
Let $|D|={g}_{8}^{3}$  be free of base points on a curve $C$  of genus nine. take a point $p\in C$  and let ${g}_{7}^{2}=|D-p|$  be. So $C$  is birational to a plane curve $\Gamma$  of geometric genus nine and degree seven with six singular points. The lines through a singular point $p$  of $\Gamma$  cut out a ${g}_{7-s}^{1}$  , where $s$  is the multiplicity of $p$  , so $C\in {\mathcal{ℳ}}_{9,7-s}^{1}\subseteq {\mathcal{ℳ}}_{9,5}^{1}$  .
In this case we have that $\mathcal{G}{\mathcal{P}}_{9}$  is a divisor in ${\mathcal{ℳ}}_{9}$  with divisorial components ${\mathcal{ℳ}}_{9,5}^{1},\mathcal{G}{\mathcal{P}}_{9,6}^{1},\mathcal{G}{\mathcal{P}}_{9,7}^{1},{\mathcal{ℳ}}_{8}^{1},\mathcal{G}{\mathcal{P}}_{9,8}^{2}$  .
3.3. For studying the locus $\mathcal{G}{\mathcal{P}}_{10}$  , we analyze for which curves $C\in {\mathcal{ℳ}}_{10}$  , the Petri map ${\mu }_{D}$  is not injective for a primitive linear series $|D|={g}_{d}^{r}$  for $2r\le d\le g-1=9$  .
3.3.1. $r=1,d=2,3,4,5,6,7,8,9$  . By gonality ${\mathcal{ℳ}}_{10,d}^{1}\subset \mathcal{G}{\mathcal{P}}_{10,6}^{1}$  for $d=2,3,4,5$  .
Since $\rho \left(10,1,6\right)=0$  , then by ([5], p.360-361) $\mathcal{G}{\mathcal{P}}_{10,6}^{1}$  is a divisorial component of $\mathcal{G}{\mathcal{P}}_{10}$  .
By theorem 2 we have that the locus of 7-gonal curves $C$  of genus ten with a $|D|={g}_{7}^{1}$  for which ${\mu }_{D}$  is not injective is the divisor $\mathcal{G}{\mathcal{P}}_{10,7}^{1}$  . By theorem 1, the locus of curves $C\in {\mathcal{ℳ}}_{10}$  with a $|D|={g}_{8}^{1}$  for which ${\mu }_{D}$  is not injective is the divisor $\mathcal{G}{\mathcal{P}}_{10,8}^{1}$  .
The locus of curves of genus ten with an autoresidual ${g}_{9}^{1}$  is the Teixidor component ${\mathcal{ℳ}}_{10}^{1}$  . So for pencils we obtain the components $\mathcal{G}{\mathcal{P}}_{10,6}^{1},\mathcal{G}{\mathcal{P}}_{10,7}^{1},\mathcal{G}{\mathcal{P}}_{10,8}^{1},{\mathcal{ℳ}}_{10}^{1}$  .
3.3.2. $r=2,d=4,5,6,7,8,9.$  Let $C$  be a curve of genus ten. If $C$  has a ${g}_{4}^{2}$  , then $C$  is hyperelliptic. By genus formula a curve $C$  of genus ten does not admit a ${g}_{5}^{2}$  . If $C$  has a ${g}_{6}^{2}$  we have several cases: If the ${g}_{6}^{2}$  is not birational, then $C\in {\mathcal{ℳ}}_{10,4}^{1}\subset \mathcal{G}{\mathcal{P}}_{10,6}^{1}$  .
Now if ${g}_{6}^{2}$  is birational, in this case the image is a smooth curve $\Gamma$  in ${\mathbb{P}}^{2}$  of degree six. Projections through a point of $\Gamma$  gives ${g}_{5}^{1}$  , so $C\in {\mathcal{ℳ}}_{10,5}^{1}\subset \mathcal{G}{\mathcal{P}}_{10,6}^{1}$  .
A curve $C$  of genus ten with a ${g}_{7}^{2}$  has a birational model in ${\mathbb{P}}^{2}$  as a curve of degree seven with $\delta =5$  singularities, some of them infinitely near. Note that the lines through a singularity of $\Gamma$  cut out a ${g}_{7-s}^{1}$  where $s$  is the multiplicity of the singularity. So $C\in {\mathcal{ℳ}}_{10,7-s}^{1}\subseteq {\mathcal{ℳ}}_{10,5}^{1}\subset \mathcal{G}{\mathcal{P}}_{10,6}^{1}$  where $2\le s\le 5$  .
Let $C$  be a curve of genus ten with a $|L|={g}_{8}^{2}$  . If the ${g}_{8}^{2}$  is not birational then we have two cases:
(a). The degree of the image ${C}_{0}$  of $C$  under the ${g}_{8}^{2}$  is two, in this case we have that $C$  has a ${g}_{4}^{1}$  , then $C\in {\mathcal{ℳ}}_{10,4}^{1}\subset \mathcal{G}{\mathcal{P}}_{10,6}^{1}$  .
(b). The image ${C}_{0}$  of $C$  under the ${g}_{8}^{2}$  has degree four, so the genus of ${C}_{0}$  is three, and we have a double cover $C\to {C}_{0}$  . In this case if we compose this double cover map with a one dimensional family of ${g}_{3}^{1}$  on ${C}_{0}$  , we have one dimensional family of ${g}_{6}^{1}$  on $C$  . Then $\text{dim}{W}_{6}^{1}\left(C\right)\ge 1$  . In thi case for a $z\in {W}_{6}^{1}\left(C\right)$  free of base points, we have that $1\le \text{dim}\left({W}_{6}^{1}\left(C\right)\right)\le {T}_{z}\left({W}_{6}^{1}\left(C\right)\right)=\rho \left(10,6,1\right)+\text{dim ker}{\mu }_{z}=\text{dim ker}{\mu }_{z}$  , so $C\in \mathcal{G}{\mathcal{P}}_{10,6}^{1}$  .
(c). When the ${g}_{8}^{2}$  is birational, we have that the image ${C}_{0}$  of $C$  under this ${g}_{8}^{2}$  is a singular curve of degree eight and genus ten with 11 singularities, some of them infinitely near. The lines through a singular point of multiplicity $s$  cut out a ${g}_{8-s}^{1}$  on $C$  . For $s\ge 3$  , $C\in \mathcal{G}{\mathcal{P}}_{10,6}^{1}$  . For $s=2$  . The lines through a double point cut out a $|D|={g}_{6}^{1}$  on $C$  . We can asumme that $|D|$  is free of base points. Denote by ${f}_{D}:C\to {\mathbb{P}}^{1}$  the map induced by $|D|$  . If $\text{ker}{\mu }_{D}\ne 0,$  then $C\in \mathcal{G}{\mathcal{P}}_{10,6}^{1}$  .
Now suposse that ${h}^{0}\left({K}_{C}\left(-2D\right)\right)=\text{dim ker}{\mu }_{D}=0$  , then ${h}^{0}\left(2D\right)=12-10+1+{h}^{0}\left({K}_{C}\left(-2D\right)\right)=3$  . $|2D|$  is free of base points and for a ramification point $y$  of ${f}_{D}$  , ${h}^{0}\left(C,2D-2y\right)={h}^{0}\left(C,2D\right)-1=2$  . Note also that $|2D|$  induces the composed map $\tau :C\stackrel{{f}_{D}}{\to }{\mathbb{P}}^{1}\stackrel{v}{\to }{\mathbb{P}}^{2}$  , where $v\left(t\right)=\left[1:t:{t}^{2}\right]$  . $\tau$  is not degenerated and since ${h}^{0}\left(C,2D-2y\right)={h}^{0}\left(C,2D\right)-1=2$  , the image $\tau \left(C\right)$  is a conic in ${\mathbb{P}}^{2}$  with a cusp at $\tau \left(y\right)\in \tau \left(C\right)$  which is impossible. So we have that $\text{ker}{\mu }_{D}\ne 0$  . So $C\in \mathcal{G}{\mathcal{P}}_{10,6}^{1}$  .
Let $C$  be a curve of genus ten with a $|D|={g}_{9}^{2}$  free of base points for which ${\mu }_{D}$  is not injective. If $|D|$  is not birational, then $|D|$  induces a ${g}_{3}^{1}$  onto a elliptic curve $E$  .
Composing this ${g}_{3}^{1}$  with the one dimensional family of ${g}_{2}^{1}$  on $E$  , we have that $C$  it has a family of positive dimension of ${g}_{6}^{1}$  . We deduce as in $\left(b\right)$  that $C\in \mathcal{G}{\mathcal{P}}_{10,6}^{1}$  .
If $|D|$  is autoresidual, then $|D|$  has finitely many base points. Let $p$  be a base point of $|D|$  , so $|D-p|={g}_{8}^{2}$  . By 3.3.2, $C\in \mathcal{G}{\mathcal{P}}_{10,6}^{1}$  .
For $|D|$  free of base points and not autoresidual, $\text{ker}{\mu }_{D}=0$  . (see [1, exercise A1, p. 196]).
In this case for linear series of dimension two we obtain the divisorial component $\mathcal{G}{\mathcal{P}}_{10,6}^{1}$  and $\mathcal{G}{\mathcal{P}}_{10,8}^{1}$  .
3.3.3. $r=3,d=6,7,8,9.$  By Castelnuovo's bound, a curve $C$  of genus $g=10$  with a ${g}_{d}^{3}$  for $d=6,7,8,9$  is not birational onto its image in ${\mathbb{P}}^{3}$  under this ${g}_{d}^{3}$  , so $C$  has a ${g}_{d}^{2}$  . By 3.3.2, for $d=6,7,8$  we have that $C\in \mathcal{G}{\mathcal{P}}_{10,6}^{1}$  . For $d=9$  , consider a $|D|={g}_{9}^{3}$  , and let $p\in C$  such that $|D-p|={g}_{8}^{2}$  , then $C\in \mathcal{G}{\mathcal{P}}_{10,6}^{1}$  by 3.3.2.
3.3.4. $r=4,d=8,9.$  By Castelnuovo's bound, a curve $C$  of genus ten satisfies $g=10>\pi \left(d,4\right)$  for $d=8,9$  , then $C$  is not birational onto its image under a ${g}_{d}^{4}$  for $d=8,9$  , so $C$  lies on a hyperplane $H\simeq {\mathbb{P}}^{3}$  in ${\mathbb{P}}^{4}$  , then by 3.3.3 for $d=8,9$  , $C\in \mathcal{G}{\mathcal{P}}_{10,8}^{1}$  .
$\mathcal{G}{\mathcal{P}}_{10}$  is a divisor with divisorial components $\mathcal{G}{\mathcal{P}}_{10,6}^{1},\mathcal{G}{\mathcal{P}}_{10,7}^{1},\mathcal{G}{\mathcal{P}}_{10,8}^{1},{\mathcal{ℳ}}_{10}^{1}$  .
The above analisys show that the divisorial components of $\mathcal{G}{\mathcal{P}}_{g}$  for $g\le 10$  are the ones that I mentioned in the introduction and the divisors that I found in section 2.
It should be interesting to find new divisorial components of $\mathcal{G}{\mathcal{P}}_{g}$  for higher genus.
For this we will need to study the geometry of special curves in ${\mathbb{P}}^{r}$  for $r\ge 3$  .
References

1. E. Arbarello, M. Cornalba, Su una congettura di Petri, Comment.Math Helvetici. (56) (1981) 1-38
2. E. Arbarello, M. Cornalba, P. Griffiths, J. Harris , Geometry of Algebraic Curves Volume I , Grundlehrem der Mathematischen Wissenschaften 267. Springer-Verlag 1984.
3. A. Castorena , Curves of genus seven that do not satisfy the Gieseker-Petri theorem, To appear in Bolletino della Unione Matematica Italiana (2005)
4. C. Ciliberto , Canonical surfaces with ${p}_{g}={p}_{a}=4$  and ${K}^{2}=5,...,10$  , Duke Mathematical Journal Vol. 48 No. 1 (1981) 121-157.
5. D. Eisenbud, J. Harris , The Kodaira dimension of the moduli space of curves of genus $g\ge 23$  , Invent. Math. 90(1987), no. 2, 359-387.
6. D. Eisenbud, J. Harris , Irreducibility of some families of linear series with Brill-Noether Number -1 , Ann.Scient.Éc.Norm.Sup. 4 série t. (22) (1989) 33-53
7. Gavril Farkas. , Gaussian maps. Gieseker-Petri loci and large Theta-Characteristics , arXiv:math.AG/0402042 v1 3 Feb 2004.
8. J. Harris, I. Morrison , Moduli of curves. Graduate texts in Mathematics 187. Springer-Verlag 1998.
9. F. Steffen , A generalized principal ideal theorem with applications to Brill-Noether theory , Invent. Math. (132) (1998) 73-89.
10. M. Teixidor i Bigas , half-canonical series on algebraic curves, Trans.Amer.Math.Soc. (302) (1987) 99-115
11. M. Teixidor i Bigas, The divisor of curves with a vanishing theta-null , Compositio Mathematica (66) (1988) 15-22

Abel Castorena Instituto de Matemáticas UNAM, Campus Morelia. Apdo. Postal 61-3(Xangari). C.P. 58089.
Morelia, Michoacán. MEXICO. e-mail address: abel@matmor.unam.mx