How many zeros does have inside ? And inside ?
No formula finds these roots — but Rouché's theorem counts them without finding them.
Why this is asked: Rouché is a counting tool: split the polynomial into a dominant term and the rest, verify the strict inequality on the circle, and read off the zero count. Watch for the extra factor when the integrand is (zf)′/(zf).
The argument principle counts zeros minus poles by watching how far the image winds, Rouché turns that count into a comparison with a simpler function, and the open mapping theorem is the structural fact underneath both.
Because "how many roots lie in this region" is a question with no elementary method and a two-line complex-analytic one, and because Rouché's hypothesis — a strict inequality holding on the whole contour — is exactly the sort of condition an option can weaken without looking different.
Let be meromorphic inside and on a simple closed contour , with no zeros or poles on . Then
the number of zeros minus poles inside, each counted with multiplicity.
The mechanism is worth seeing once. Near a zero of order , with , so
and the second term is holomorphic. So has a simple pole with residue exactly at every zero of order , and residue at every pole of order . The residue theorem does the rest.
The name comes from the other reading: is the total change in , whose imaginary part is the change in . Dividing by gives the number of times winds around the origin. Counting zeros and watching the image wind are the same computation.
If everywhere on , then and have the same number of zeros inside.
The picture is the proof: walking a dog around a lamp post, on a lead shorter than the distance to the post, the dog circles the post exactly as often as you do. is the walker, the dog, the lead.
Two hypotheses are doing work and both get removed in options.
Strict inequality. is not enough: take and on , where and has infinitely many zeros.
On the whole contour. The inequality must hold at every point of , not merely somewhere or on average.
The standard use is bracketing roots by annulus. For :
So four roots sit in the annulus . Notice the dominant term changes between the two circles — choosing which term plays is the whole skill.
Rouché also gives the fundamental theorem of algebra: on a large enough circle the leading term dominates the rest, so a degree- polynomial has exactly roots.
A non-constant holomorphic function on a domain maps open sets to open sets.
This is why can have no interior maximum: the image of a neighbourhood of is a neighbourhood of , so it contains points of larger modulus. The maximum modulus principle is a corollary rather than a separate fact.
Two more consequences:
A real-variable contrast worth carrying: maps the open interval onto , which is not open. Openness of the image is a complex phenomenon.
" on , so Rouché applies." Strict. , gives . The weakened inequality is the commonest way this option is made wrong.
"Rouché needs inside the contour." On the contour only. Requiring it inside would make the theorem useless, since that is where the zeros are.
"The argument principle counts zeros." Zeros minus poles. For a meromorphic function the two cancel, and an option will supply a function with poles and ask for the winding number.
"Multiplicity does not matter." Everything here is counted with multiplicity. has one zero as a set and three as a count, and the integral returns three.
" implies is injective." Only locally. has non-vanishing derivative everywhere and is not injective on — the same local-versus-global gap as the inverse function theorem.
" injective implies ." This one is true for holomorphic , and is worth recognising among false options. It fails for real functions: is injective on with .
"A non-constant holomorphic map can have a closed image." Its image is open. A non-constant holomorphic function on a domain never maps onto a closed set like , and never takes only real values.
See it move
The trap here
“f′(z) ≠ 0 everywhere implies f is injective” — false
f(z) = eᶻ on
f′ = eᶻ never vanishes, yet . Non-vanishing derivative gives only local injectivity.
Check yourself
How many roots does the polynomial have in the open disc { : |z| < 1}?
More in Zeros and Mappings
Take it further
Last revised . Found a mistake? Tell us — corrections are the fastest thing we act on.
Next: Conformal maps, Möbius transformations, Schwarz lemma
Create a free account to keep your place and have this feed your study plan.
Open this in the full syllabus view · Unit 2