474 Chapter^7
have, by (7.12-2),
_1 J J(()d( = _1 J J(()d(
27l"i ( - z 27ri ( - z
C C1
= _1 j f(z)d( + _1 j f(() - f(z) d(
27l"i ( - z 21l"i ( - z
C1 C1
= nc(z)f(z) + ~ J !(() - f(z) d(
27ri ( - z
(7.17-2)
C1
But the last integral vanishes. In fact, if we let k = nc 1 (z), we have
~ j J(() -J(z) d( ~ _2._ · ~ · 2JkJ7rr = JkJ1:
27l"i ( - z 211". 7'
C1
and since E > 0 can be taken arbitrarily small, it follows that the integral
above is zero.
The reader will note that the left-hand side of (7.17-2), and the term
nc(z)f(z) on the right-hand side, are independent of r = r(1:). Hence
formula (7.17-1) holds.
Alternatively, we may prove that
J
J(()-J(z)d(=O
(-z
C1
by observing that the integrand is analytic on and within Ci, except at
the poinLz where it is undefined. However, since
f(()-f(z) =f'(z)+TJ
(-z
where 7J ---+ 0 as ( --+ z, we see that the left-hand side is locally bounded
at z, and it suffices to apply Theorem 7.21.
Next, suppose that z E Ext C*. In this case nc(z) = 0 and
J
J(()d( = 0
(-z
c
also, since the integrand is now analytic in and within C*, so that (7.17-1)
holds good for this case.
Note Formula ( 7 .17-1) is still valid if f is analytic in D = Int C* and