Page 200 -
P. 200

§4. The Differential Equation for ℘(z)  177

        L, and by (2.5) there are three zeros z 1 , z 2 ,and z 3 of f (z) with z 1 + z 2 + z 3 ∈ L.
        Thus under h −1  the three intersection points of y = λx + ν, which add to zero on E,
        are transformed to three points z 1 + L, z 2 + L,and z 3 + L ∈ C/L which add to zero
        on C/L. This proves the theorem.
           Using the above considerations in proving h −1  is a group morphism and the

        techniques of 1, §2, we can derive the addition formulas for ℘(z).For ℘ (z i ) =
        λ℘(z i ) + ν, where i = 1, 2, 3, the equation
                              3          2
                            4x − (λx + ν) − g 2 x − g 3 = 0
                                                                    2
        has three roots, and the sum of the roots is related to the coefficient of x by the
        relation
                             λ 2
                                = ℘(z 1 ) + ℘(z 2 ) + ℘ (z 3 ) .
                              4
        Given z 1 , z 2 , choose z 3 =−(z 1 + z 2 ) mod L, and for ℘(z 1 )  = ℘(z 2 ),wehave


        ℘ (z 1 ) − ℘ (z 2 ) = λ(℘(z 1 ) − ℘(z 2 )). Since ℘ is an even function, this gives the
        following first division formula. The second formula is obtained by letting z 1 and z 2
        approach z in the limit so z 3 becomes −2z.
        (4.4) Addition Formulas for the Weierstrass ℘-Function.
                                                                  2




                                                1  ℘ (z 1 ) − ℘ (z 2 )
         (1)        ℘ (z 1 + z 2 ) =−℘(z 1 ) − ℘(z 2 ) +           .
                                                4   ℘(z 1 ) − ℘(z 2 )
                                                        2

                                              1   ℘ (z)


         (2)                 ℘(2z) =−2℘(z) +             .
                                              4   ℘ (z)

           In 3(2.3) we introduced the notion of an admissible change of variable carrying
        one normal form of the equation of an elliptic curve into another or equivalently
        defining an isomorphism between two elliptic curves.
        (4.5) Remark. An admissible change of variable defining an isomorphism f :
                                 2
                                                           2
                                                                 3
                                       3
        E → E, where E is given by y = 4x −g 2 x −g 3 and E by y = 4x −g x −g ,




                                                                    2     3
                                      3
                         2
        has the form xf = u x and yf = u y . Moreover, we have by 2(2.4) the relations
        g 2 = u g and g 3 = u g .
              4
                          6
                2           3

           In (1.3) we considered the equivalence between two lattices L and L together
        with the corresponding isomorphism defined between their related tori T = C/L


        and T = C/L . Under this isomorphism there are transformation relations for the
        Weierstrass ℘-function and its related coefficient functions.
        (4.6) Proposition. For two equivalent lattices L and L = λL the following rela-

        tions hold:
                              2
                                                     3
                    ℘(z, L) = λ ℘(λz,λL),  ℘ (z, L) = λ ℘ (λz,λL),


                                         4
                                                                 6
                     2k
           G 2k (L) = λ G 2k (λL),  g 2 (L) = λ g 2 (λL),  and  g 3 (L) = λ g 3 (λL).
   195   196   197   198   199   200   201   202   203   204   205