20 2Basics
where it is understood that the 1 on the right hand side stands for the unit matrix.
Comparison of (2.32) and (2.33) with (2.26) and (2.25) reveals: the inverseU−^1 of
the orthogonal matrixUis just its transposedUT:
U−^1 =UT, (2.34)or
Uμν−^1 =Uνμ. (2.35)
Use of the inverse transformation in considerations similar to those which lead to
(2.31) and of (2.35) yield the orthogonality relation with the summation index at the
back,
UμλUνλ=δμν, (2.36)
or, equivalently,
U·UT= 1. (2.37)
Summary
The coordinate transformation
rμ′=Uμνrν, (2.38)
where the matrixUμνhas the property
UμλUνλ=UλμUλν=δμν (2.39)guarantees that the scalar product of two vectors (2.8) and consequently, the expres-
sion (2.4) for the length of a vector are invariant under this transformation. Further-
more, the relation (2.39) means that the reciprocalU−^1 ofUis equal to the transposed
matrixUTwhich, in turn is defined byUμνT =Uνμ.
Simple Examples
The simplest examples for transformation matrices which obey (2.39)areUμν=δμν
andUμν=−δμν, or in matrix notation:
U=δ:=⎛
⎝
100
010
001
⎞
⎠, U=−δ:=⎛
⎝
−10 0
0 − 10
00 − 1
⎞
⎠, (2.40)
which, respectively, induce the identity transformation and a reversal of the directions
of the coordinate axes. The latter case means a transformation to the ‘mirrored’
coordinate system.