A map that bends every line but keeps every angle
Here is a map of the plane that bends almost everything. Take . Feed it the horizontal and vertical grid lines of the -plane — the ordinary graph-paper grid — and watch what comes out in the -plane. The straight lines become parabolas. The square cells bulge and stretch, small near the origin, enormous far away. Nothing is straight any more. Nothing is the same size.
And yet, almost everywhere, one thing survives untouched: the angles. Where two grid lines crossed at a right angle in the -plane, their bent images still cross at a right angle in the -plane. The map throws away straightness and throws away scale, but it refuses to throw away angle.
That refusal has a name. A map that preserves angles is called conformal, and the surprising fact of this chapter is that every analytic function is conformal — wherever its derivative isn't zero.
You already know why, you just haven't connected it yet.
In Chapter 1 you learned that multiplying by a complex number does exactly two things: it scales by and rotates by . In Chapter 3 you learned that an analytic function has a well-defined derivative — a single complex number attached to each point. Put those together. Up close, near a point , an analytic function does almost nothing except multiply small displacements by . And multiplying by a complex number rotates and scales.
Rotate-and-scale is the gentlest thing a map can do to a neighbourhood. It can spin the picture and blow it up or shrink it, but it cannot shear, cannot reflect, cannot distort one direction more than another. So the angle between two little arrows leaving is preserved: both arrows get spun by the same angle , and the angle between them never moves.
This is the chapter where the geometry of Unit 1 pays off, and it pays off twice. First, immediately: Chapter 6 is about Möbius transformations, the single most useful family of conformal maps, and everything there rests on the angle-preservation idea you build here.
Second, in applications — this is the CO-4 thread. Because conformal maps preserve angles, they also preserve the equation that governs steady heat flow, electrostatics, and ideal fluid flow: Laplace's equation. A harmonic function (Chapter 4) stays harmonic after a conformal change of variables. So a hard boundary-value problem on an awkward region can be conformally mapped to the same problem on a simple region — a disc, a half-plane — solved there, and mapped back. One idea, "analytic maps preserve angles," is at once the gateway to Chapter 6 and the reason complex analysis turns up in heat-transfer and field-theory courses two semesters from now.
What "preserves angles" actually means
A map is a picture of one plane drawn on another
Write . Think of two copies of the plane side by side: the -plane on the left, the -plane on the right. The function takes each point on the left to a point on the right. A curve on the left becomes a curve on the right; a region becomes a region. The whole subject of conformal mapping is the study of what shapes go to what shapes under a given .
This reframes everything from Chapters 3–4. " is analytic" stops being an algebraic property of and and becomes a geometric one: it controls how the left picture is redrawn on the right.
The angle between two curves
Take two smooth curves and that cross at a point . At the crossing, each curve has a tangent direction; the angle between the curves is the angle between those two tangents, measured with a sign (counterclockwise from to ). Under the curves become images and crossing at , with their own angle between them.
Conformal at means: for every such pair of curves through , the angle between the images equals the angle between the originals — same size, same sense (same direction of turning).
The "same sense" clause is the part students drop, and it is exactly what separates a conformal map from its mirror-image cousin. We name that cousin in Beat 3.
Why scaling and rotating cannot disturb an angle
Here is the picture under the surface, the one Beat 1 promised. Near , the map sends a tiny displacement to a tiny displacement . Writing :
Every little arrow leaving is rotated by the same and scaled by the same . Rotating all arrows by a common angle leaves the angles between them unchanged; scaling them all by a common factor does too. So angles survive — provided , because only then is and a sensible, single direction to rotate by.
When , the arrow collapses, the linear approximation says only "," and the whole argument falls silent. Those points are special, and they are where conformality breaks. They have a name too, coming up next.
Analytic ⟹ conformal, except where the derivative dies
Let be analytic on a domain , and let with . Then the mapping is conformal at : it preserves both the magnitude and the sense of angles between curves through .
The proof is the Beat 2 argument made precise: near , , a multiplication by the fixed nonzero number ; multiplication rotates every tangent by and scales by , so the angle between any two tangents is unchanged in size and sense. (Reveal 2 walks this through in three steps — it is a clean exam-grade proof.)
Critical points — where conformality fails
A point where is analytic but is called a critical point of . At a critical point the mapping is not conformal.
Critical points are the zeros of , so for any non-constant analytic they are isolated — scattered, never filling a region. At an ordinary point, angles are preserved; at a critical point, they are not merely disturbed but multiplied.
The precise rule: if and the first non-vanishing derivative at is the -th one, then curves crossing at at angle are mapped to curves crossing at . Angles are multiplied by at a critical point. For the derivative vanishes only at , where : angles double there — the one case you'll see worked here, with the general as its natural extension. That doubling is the one place the grid betrays its right angles — at the origin, where a quarter-turn becomes a half-turn.
Conformal, isogonal, anti-conformal
Isogonal = preserves the magnitude of angles, sense or not. A conformal map is an isogonal map that also preserves sense (so conformal maps are the sense-preserving isogonal maps). An anti-conformal map is an isogonal map that reverses sense. The cleanest anti-conformal example is — a reflection across the real axis: same angle sizes, flipped direction, so it is isogonal but not conformal. And isn't analytic (fails Cauchy–Riemann), which is exactly why it isn't conformal — conformality is the analytic, sense-preserving member of the isogonal family.
Try it: deform the grid, watch the angles
Three elementary maps, for reference as you play:
Linear map (). Derivative everywhere, never zero, so conformal at every point. Geometrically it is pure rotate-scale-translate: rotate by , scale by , shift by . It preserves angles and shapes (a similarity); only position, size and orientation change. This is the "nothing surprising happens" baseline.
Square map . Derivative , zero only at . Conformal everywhere except the origin; angles double at the origin. In polar form : radii get squared, angles get doubled. A ray at angle maps to a ray at angle , which is why the first quadrant () opens out to the entire upper half-plane ().
Inversion . Derivative , which is never zero on the map's domain. So is the instructive contrast to : it is conformal at every point where it is defined — it has no critical points at all, whereas has exactly one. And one caution that reliably trips students: the origin is not a critical point of . A critical point is a place where is analytic and vanishes; at the function is simply undefined — it isn't in the domain, so it can't be a critical point. ("Undefined here" and "critical point here" are different statements.) What does to circles and lines is the heart of Chapter 6; for this chapter it is enough to see that it bends the grid while preserving every angle it touches.
Notice what is doing in the angle-probe panel. Its argument is exactly how much the picture is rotated at ; its modulus is exactly how much the picture is magnified there. The derivative you computed back in Chapter 3 as a limit of difference quotients turns out to be a little instruction sheet: "at this point, spin by and zoom by ." That is what a complex derivative is, geometrically.
And the one point where the instruction sheet is blank — , "spin by nothing, zoom by nothing" — is precisely the point where angles stop being preserved. No rotation angle is defined, so there is no common angle to protect.
Test yourself before moving on
Commit to an answer before revealing anything. Recognising the right answer once you see it feels like knowing it. It isn't the same thing.
For the map , at which of these points does the mapping fail to be conformal?
A map is conformal at , with . Two smooth curves cross at at an angle of 40°. At what angle do their images cross at ?
Under , two rays leave the origin with 30° between them. What is the angle between their images at ?
Parametrise the line by : a point on it is . Then
So writing ,
Eliminate the parameter: from the second equation , and substituting,
This is a parabola opening in the negative- direction, with vertex at .
Check one point. At : , , so . The formula gives . ✓
Horizontal lines with map to parabolas opening the other way (). The degenerate case is the real axis, which maps to the non-negative real axis — a ray, not a parabola. Away from the origin the two parabola families are mutually orthogonal — they cross at right angles — which is exactly what the conformality of guarantees. Same orthogonal-families idea as Chapter 4, now produced by a mapping rather than by a harmonic conjugate.
Prove: an analytic function is conformal at any point where . Three sub-steps; attempt each on paper before revealing.
Differentiability at means
where the error shrinks faster than as . So for a small displacement , the image displacement is
To leading order, the map is multiplication by the fixed number .
From Chapter 1, multiplying by scales modulus by and adds to the argument:
Every small displacement leaving , whatever its direction, is rotated by the same angle . Here is where is essential: only a nonzero number has a well-defined argument , so only then is "rotate by " a meaningful instruction.
By Step 2 each tangent direction is rotated by :
The angle between the image curves is the difference of image directions:
the original angle — unchanged in size, and unchanged in sign, so unchanged in sense. The map is conformal at . ∎
The proof also explains the exception for free. If , Step 1 leaves no linear term and Steps 2–3 never start — which is exactly why critical points are not conformal. One argument delivers both the theorem and its boundary.
How this concept appears on exams
In standard Indian engineering math exams, and likely in MCC201A unless your class notes differ, conformal mapping shows up in three forms.
CO-1 form — define and identify. "Define a conformal mapping" / "Distinguish conformal and isogonal mappings" / "Define a critical point and find the critical points of ." Roughly 3–5 marks; budget 4 minutes. The definition must include both magnitude and sense of angles, or it reads as the bare isogonal definition and loses a mark. For critical points, just solve .
CO-1 / CO-4 form — discuss a mapping. "Discuss the transformation " or "Find the image of the region in the first quadrant between and under ." Roughly 6–8 marks; budget 8 minutes. The polar reading does most of the work; state where the map is and isn't conformal.
CO-1 form — justify the theorem. "Show that an analytic function is conformal where ," or "Prove that the angle of intersection is preserved." Roughly 5–7 marks; budget 6 minutes. Reveal 2 is the model answer — the three-line linearisation , then "multiplication rotates all tangents equally."
Three patterns that commonly cost marks on this material.
Dropping "sense" from the definition. Writing "a conformal map preserves the magnitude of angles" and stopping. That is the isogonal definition — the larger family. Conformal preserves magnitude and sense; the missing clause is the whole distinction, and examiners test exactly that boundary with .
Forgetting that critical points are not conformal. A question may hand you a map and a point that happens to be a critical point and ask about angle preservation, expecting you to notice and say "not conformal here; angles are multiplied." Treating every analytic map as conformal everywhere walks straight into it.
Confusing "conformal" with "shape-preserving." Conformal preserves angles, not shapes or sizes. A conformal map can bend a straight line into a parabola (you did this in Reveal 1) and still be perfectly conformal. Only similarity maps () preserve shape; the general conformal map preserves angle alone.
Six things to carry out of this chapter
- A mapping redraws the -plane onto the -plane. Curves go to curves, regions to regions; the subject is which shapes go where.
- Conformal = preserves angles in magnitude and sense. The sense clause is non-negotiable.
- Every analytic function is conformal wherever . The reason: near the map is multiplication by , which rotates all tangents by and scales by , leaving angles between them fixed.
- is the local instruction sheet. Its argument is the local rotation, its modulus the local magnification. The derivative is the rotate-and-scale.
- Critical points are zeros of , and conformality fails there. Angles are multiplied by (the order of the first non-vanishing derivative). For , angles double at .
- Isogonal = angle magnitude preserved (sense not necessarily). Conformal maps are the sense-preserving isogonal maps; anti-conformal maps (e.g. , not analytic) reverse sense.
What comes next. Chapter 6, Möbius transformations, takes the single most important family of conformal maps — — and works out exactly what they do to circles, lines and regions. The angle-preservation reflex you built here is the foundation; Möbius maps are conformal everywhere they're defined, and their special magic (circles and lines go to circles and lines) sits on top of that.
Before you open Chapter 6: take two or three maps — , , — and for each, find the critical points (solve ) and sketch the image of one simple curve. ( wasn't in the playground, but the same rule applies — find and check whether it ever vanishes.) The "discuss the transformation" exercises in Zill, Complex Analysis, Ch. 7 [verify the section against your copy] are the highest-value drill. Don't re-read this chapter; produce the pictures yourself.
Same material, another voice
If a different explanation would help, this one is worth your time — free, from MIT OpenCourseWare:
- Watch: Herbert Gross, Calculus Revisited, Lecture 3 — Conformal Mappings — a geometric chalkboard introduction to conformal maps, this chapter in motion.
✓ Chapter complete. Your progress, and every quiz answer, is saved on this computer — revisit any time.