Prym-Tyurin varieties coming from correspondences with fixed points (Q877708)

Fix a smooth projective curve \(C\) over the complex number field. A correspondence is a line bundle \(L\) on \(C \times C\); the correspondence is effective if it is defined by an effective divisor on \( C \times C\). The author considers in this paper effective correspondences which can admit fixed points. Let \(p \in C\) and define the line bundle \( L(p) = L|_{ p \times C}\). The author's main theorem is: Theorem 1.1. Let \(L\) be an effective fixed symmetric correspondence on \(C \times C\) of bidegree \((d,d)\) with \( 2 n \) fixed points and \( \gamma_L\) the endomorphism of the Jacobian \(J\) induced by \(L\).~Suppose \( n \leq d\) and a) \({\gamma}_L\) satisfies \( (1 - \gamma)( \gamma + q -1) = 0 \) for \(q \geq 2\). b) There are \(n \) fixed points \( p_1, \ldots, p_n\) such that \( p_1, \ldots, p_i \in D_i \) and \( p_i \notin D_i - p_i \) for \( i=1, \ldots, n \) where \( L(p_i) = O_C( D_i)\) and \( D_i\) are effective divisors. Defining \( P= \text{Im} ( 1-\gamma) \) is a Prym-Tyurin (PT) variety of exponent \(q\) for the curve \(C\). Moreover, there exists theta divisors \( \Theta , \Psi\) on \(J\) and \(P\) respectively such that \( \iota^{\ast} \Theta = q \Psi\). In the case that \(\gamma\) is an endomorphism induced by a fixed point free symmetric correspondence satisfying the equation given in a) has been proven by \textit{V. Kanev} [Compos. Math. 64, No. 3, 243--270 (1987; Zbl 0694.14009)]. The simplest case of a correspondence with fixed points was already considered by \textit{D. Mumford} [in: Contributions to Analysis, 325--350 (1974; Zbl 0299.14018)] where he studies the case of Prym varieties \((q= 2\)) associated to a double cover over a curve with two branch points. The paper is organized as follows. In section one the main theorem of the paper is stated. In section two the author recalls some generalities about correspondences and abelian varieties.~ In section three the author then states and proves proposition 3.1 which states that there exist theta divisors \(\Theta, \Psi_0\) on \(J \) and on a p.p.a.v. \(P_0\) such that \( j^{\ast} \Theta = q \Psi_0 \).~The proof of theorem 1.1 then follows from this proposition and proposition 2.1. ~Lemma 3.1 is then proven and used in the proof of proposition 3.1. In section four the author considers some examples. The first example is that of a covering of degree 3 with two ramification points with base a hyperelliptic curve of genus greater than equal to three with the hyperelliptic involution and a covering of \(P_1\) given by the linear system \( g_2^1\).~ Using this linear system a new curve \(C\) and an effective symmetric correspondence \(D\) is constructed on \(C\) of bidegree \(4\) assuming \(C\) smooth and irreducible. If \(\gamma\) is the associated correspondence the author verifies conditions a) and b) of theorem 1.1 to prove that \(P= \text{Im}(1 -\gamma)\) is a PT variety of exponent \(3\) for the curve \(C\). The author does not prove the irreducibility of \(C\) but refers to \textit{H. Lange, S. Recillas} and \textit{A.M. Rojas} [J. Algebra 289, No. 2, 594--613 (2004; Zbl 1089.14005)] for a similar argument to prove it.~In lemma~4.1 the author proves that \(C\) is smooth. The next examples are of coverings of \(P_1\). Namely if a \( g_{n+2}^1\) is the complete linear system defining the covering for \( n \geq 2 \) and then a curve \(C\) is defined using this linear system and an effective symmetric correspondence \(D\) on \(C\). Assuming that \(D\) is fixed point free the author uses proposition 3.1 to prove that \(P= \text{Im}( 1-\gamma)\) is a PT variety of exponent \(n \) for the curve \(C\).~On the other hand, if \(D\) has fixed points the author considers the following cases as applications of theorem 1.1. \(n=2\).~\(P = \text{Im}( 1 - \gamma) \) is a PT variety with \(q=2 \). \(n=3\).~\(P\) is a PT variety with \(q = 3\) . \(n=4\).~\(P\) is a PT variety with \( q= 4\). The author then proves in proposition 4.1 that \( P \simeq JX\) (as p.p.a.v.s) for all the above examples, first proving lemma 4.2 and then proposition 4.1, difficult to understand because of some missing notations. Namely, in the first line of the proof \( Z\) is in fact \(P\) and then after equation 4.10 \( 2_{\ast}\) should be \( 2_{\ast J(X)}\) instead. The method of proof of this lemma is very similar to that of \textit{S. Recillas}' theorem [Bol.~Soc.~Mat.~Mexicana 19, 325--350 (1974; Zbl 0343.14012)]. The paper concludes the proof of proposition 4.1 by proving that \( \alpha_n(C)\) is algebraically equivalent to \(nX\).