Page 184 -
P. 184

§2. Primitive Descent Formalism  161


                                H 1  Q, ϕ E Q  = Hom Gal Q/Q , Z/2Z
                  ϕ

            E(Q) → E (Q)
                                      ∗ 2
                                 ∗
                                Q / (Q ) .
        The homomorphism α is given in 4(5.6) and considered further in 6(3.1) in this
        context, and the homomorphism δ is defined in 7(4.2).
                           2
                                      3
           Again let x = m/e and y = n/e be the coordinates of (x, y) on E[a, b] with
        rational coefficients in lowest terms. For n = 0wehave x values given by roots of
                                                             ∗2
           2
        x(x + ax + b) = 0, and for x = 0 the value α(0, 0) = b(mod Q ), but otherwise


        there is a second root x with xx = b.
           For n  = 0wehave m  = 0 and are led to the study of the congruence properties
        of m, n, and e. From the equation

                     2
                                2 2
                          3
                                                        2
                                         4
                                                 2
                    n = m + am e + bme = m m + ame + be      4
        we introduce the greatest common divisor b 1 = (m, b) and factor m = b 1 m 1 and
        b 1 b 2 = b where (m 1 , b 2 ) = 1. The equation becomes

                                                2
                                    2
                          2
                                       2
                         n = b 1 m 1 b m + ab 1 m 1 e + b 1 b 2 e 4
                                    1  1

                               2
                                       2
                                              2
                           = b m 1 b 1 m + am 1 e + b 2 e 4  .
                                       1
                               1
                      2
              2
        Thus b divides n and so b 1 divides n giving a factorization n = b 1 n 1 . Putting this
              1
                                           2
        relation into the equation and dividing out b , we obtain the relation
                                           1

                                               2
                                       2
                            2
                           n = m 1 b 1 m + am 1 e + b 2 e 4  ,
                            1          1
                                            2
                                                                       2
                                                  4
                                    2
        where the two terms m 1 and b 1 m + am 1 e + b 2 e in the factorization of n are
                                    1                                  1
        relatively prime since (b 2 , m 1 ) = 1and (e, m 1 ) = 1.
           We can always choose the sign of b 1 so that m 1 is positive, and the above form of
                                                            2
        the equation implies that m 1 is a square which we write m 1 = M . Then M divides
        n 1 from the above equation, and we can write n 1 = MN, with all factorizations
        taking place in the integers Z.
        (2.3) Assertion. We can summarize the above discussion for a rational point (x, y)
        on E[a, b]with y  = 0 in terms of a representation as a quotient of integers in the
        form
                                 b 1 M 2        b 1 MN
                             x =        and  y =      .
                                  e 2             e 3
        For x and y reduced to lowest terms we have (M, e) = (N, e) = (b 1 , e) = 1. Also,
        since (b 2 , m 1 ) = 1, we have (b 2 , M) = (M, N) = 1. The equation for M and N is
                               2      4      2 2     4
                              N = b 1 M + aM e + b 2 e .
   179   180   181   182   183   184   185   186   187   188   189