Page 89 - Curvature and Homology
P. 89

whenever the integral converges as will always be the case in the sequel.
       (It is assumed that M is orientable and that an orientation of M has been
       chosen).  The scalar product evidently has the properties:







       where a, a,,  a,,  P, /I1 and P2 have the same degree.
         If  (a, P) = 0, a and P are said to be orthogonal.
         It should be remarked that the r operation is an isomorphism between
       the spaces  AP  (Tg) and  A"-P(T,*) at each point P of  M.


                 2.8.  Harmonic forms.  The operators  6  and  A
         There  are  several  well-known  examples  from  classical  physics
       (potential theory) where relations analogous to Laplace's  equation hold.
       The electrical potential due to a system of charges or the vector potential
       due to a system of  currents is not uniquely  determined.  To the former
       an arbitrary constant may be added and to the latter an arbitrary vector
       with  vanishing  curl.  In  defining  electrical  potential  we  begin  with  a
       vector  field  E representing the  electrical  intensity  which  satisfies the
       equation  curl  E = 0. This is the  condition  that  the  electric  field  be
       conservative.  A function f is then defined as follows:




       where r denotes  the  position  vector  of  a  point in E3 and the . denotes
       the inner  product  of  vectors  in ES. It follows  that  E = grad f and f
       is determined to within an additive constant.
         In defining the  vector  potential,  on  the  other  hand,  we  begin  with
       the magnetic induction B which satisfies the equation div B = 0. As it
       turns  out,  this  is  a  sufficient  condition  for  the  existence  of  a  vector
       field  A  (unique  up  to  a  vector  field  whose  curl  vanishes)  satisfying
       B = curl A.
         We now  re-write  the above equations as tensor equations in E3. We
       may  distinguish  between  covariant  and  contravariant  tensor  fields
       provided  the  coordinate  system  is  not  Euclidean.  Let  Ei denote  the
       components  of  the covariant vector field E and Bi the components of
       the contravariant vector  field B.  Then,
                              D, E, - D, E,  = 0               (2.8.2)
   84   85   86   87   88   89   90   91   92   93   94