Elementary Functions 259
where A = a1 + ia2, B = a1 + ia 2 , C = 2(c 1 + ic 2 ), N = 2c 3. Then
we find that
1
a1 + a1 a1 -a1 C1
D=- a2 +a2 a2 -a2 C2
4i
M+M M-M C3
1
a1 C111 C1
= 2i a2 lll2 C2
M M C3
1
a1 + ia2 a1 + ia2 c1 + ic2
= 4 ii1 - iii2 C111 - ZC\12 C1 - ic2
M M C3
A B c
1
= - f3 A () = ~E
8
M M N
8
Thus D # 0 implies that E # O, and conversely.
A generalized bilinear transformation may have either no fixed point or
one, two, three, four, or infinitely many fixed points lying on a generalized
circle. The analysis of the different cases is somewhat involved, and we
must refer the reader for this and other details to J. A. Cabri [4].
Exercises 5.2
- Let S be the set of all conjugate bilinear transformations. An iden-
tity in Sis a transformation V E S such that (VVU)(z) = U(z) and
(UVV)(z) = U(z). Show that any conjugate transformation of the form
V(z) = ~z_+ i~
Z"'(Z +a
with a complex, (3 and 1 reals, and aii + (31 = 1 is an identity in S.
- Find the fixed points of the transformation w = az + b (a # 0).
- Show that the inverse of a nonsingular bilinear transformation of the
form (5.13-3) is given by
(AN - CM)w +(CM - BN)w +(BC - AC)
z = - --
(BM - AM)w +(BM - AM)w +(AA - BB)
- Find the fixed points of w = Az + Bz + C, IAI # IBI # 0.