Two Dimensional Example: Moving Complex Axis on a Sphere Surface



First start with a complex numbers on a plane

Represent the vector V as a complex number relative to a complex coordinate axis: V=Vr+i Vi. The basis vectors are the unit real 1 represented as I , and the unit imaginary number i. Rotate the I-i axis by some angle. We can represent the rotation with a complex number of unit length. The coordinates of the new (primed)axis relative to the old are I'= I. Note that the components of V transform oppositely from the basis vectors.

The components transform contravariantly to the way the basis vectors transform. Unlike V , the rotation vector  is not a real object; it is derived from the coordinate transform. For example the rotation vector effects another rotation on the primed coordinate system is certainly not the  pictured above: that omega effects no rotation in the I'-i' system. This is very different from the behavior of the vector V which is the same vector regardless of the coordinate system. The transformation behavior of  is analogous to the Christoffel symbols of classical tensor analysis.

Next we draw the complex axis on a sphere surface. The axes may be thought of as geodesics. The displacements are shown here greatly exaggerated to demonstrate the effect of the surface curvature: One should imagine the displacements as infintesssemal.





A complex axis moves along the surface of a sphere of radius a. The axis will rotate so that the imaginary axis t lines up with the coordinate lines of ; the real axis lines up with is the geographic latitude ( equator=0 radian) and is the longitude. The scaling differential form is:

The exterior derivative of 

We now seek a one form  that solves:

is easily found to be:

The curvature 2-form  is

is a small complex number proportional to the product representing a parallel transversal about a little rectangle with sides . We now scale the little rectangle in actual distance units:

The coefficient 1/a2 is indeed the scalar curvature.

Back
Table of Contents
 
 

Rev 5/7/0 Copyright ©2000 George Raetz <bestwork.1@pcisys.net>