Page 382 - DMTH402_COMPLEX_ANALYSIS_AND_DIFFERENTIAL_GEOMETRY
P. 382

Unit 30: Geodesic Curvature and Christoffel Symbols




          This perpendicular distance can be expressed as                                       Notes
                                                         .
                                  (u, v) = (X(u, v) – X(u , v )) N (u ,v ) .
                                                    0
                                                      0
                                                            0 0
          However, since X is at least C -continuous, by Taylor’s formula, in a neighborhood of (u , v ),
                                  3
                                                                                  0
                                                                                     0
          we can write
                                              1
          X(u, v) = X(u , v ) + X (u – u ) + X (v – v ) +  2  (X (u – u )  + 2X (u – u )(v – v ) + X (v – v ) ) +
                                                        2
                                                                                    2
                                 0
                            u
                       0
                                                             uv
                     0
                                                                   0
                                                                             vv
                                                                                  0
                                                 uu
                                                       0
                                          0
                                                                        0
                                     v
                    ((u – u )  + (v – v ) )h (u, v),
                                2
                        2
                       0
                                  1
                               0
          where lim (u,v)  (u ,v )  h (u, v) = 0.
                           1
                       0 0
          However, recall that X  and X  are really evaluated at (u , v ) (and so are X , X , and X ), and so,
                                 v
                                                                    uu
                                                      0
                                                                              vv
                                                                       u,v
                                                        0
                            u
          they are orthogonal to N (u ,v ) .
                               0 0
          From this, dotting with N (u ,v ) , we get
                                0 0
               (u, v) =  (L(u – u ) + 2M(u – u )(v – v ) + N(v – v ) ) + ((u – u ) + (v – v ) )h(u, v),
                                                                          2
                                                        2
                                                                  2
                              2
                                                                0
                                        0
                                              0
                                                                         0
                             0
                                                       0
          where lim (u,v)  (u ,v )  h(u, v) = 0.
                       0 0
          Therefore, we get another interpretation of the second fundamental form as a way of measuring
          the deviation from the tangent plane.
          For  small enough, and in a neighborhood of (u , v ) small enough, the set of points X(u, v) on
                                                    0
                                                 0
                                    1
                                       2
          the surface such that (u, v) =     will look like portions of the curves of equation
                                    2
                           1        2                         2   1  2 .
                                             0
                                                            0
                                                   0
                                  0
                           2 (L(u – u ) + 2M(u – u )(v – v ) + N(v – v ) )   
                                                                  2
          Letting u – u  = x and v – v  = y, these curves are defined by the equations
                    0
                                 0
                                       Lx  + 2Mxy + Ny  = ±1.
                                                     2
                                         2
          These curves are called the Dupin indicatrix.
                                                                 
          It is more convenient to switch to an orthonormal basis where  e  and  e  are eigenvectors of the
                                                                   2
                                                             1
          Gauss map dN (u ,v ) .
                       0 0
          If so, it is immediately seen that
                                    Lx  + 2Mxy + Ny  = k x  + k y , 2
                                                  2
                                                       2
                                      2
                                                          2
                                                     1
          where k  and k  are the principal curvatures. Thus, the equation of the Dupin indicatrix is
                      2
                 1
                                          k x  + k y  = ±1.
                                            2
                                                  2
                                           1
                                                2
          There are several cases, depending on the sign of k k  = K, i.e., depending on the sign of LN – M .
                                                                                     2
                                                  1 2
          (1) If LN – M  > 0, then k  and k  have the same sign. This is the case of an elliptic point.
                     2
                              1
                                    2
          If k   k , and k  > 0 and k  > 0, we get the ellipse of equation
                               2
             1
                2
                      1
                                           x 2    y 2   1,
                                            1    1
                                           k 1   k 2
                                           LOVELY PROFESSIONAL UNIVERSITY                                  375
   377   378   379   380   381   382   383   384   385   386   387