Branch of a multivalued function Created 2026-10-03 Updated 2026-10-14
A branch of a multivalued analytic expression is a single-valued holomorphic function on a specified domain that consistently selects one of the expression's possible values. A branch cut is a convenient way to choose such a domain, rather than an intrinsic locus of singularities. On , the principal logarithm supplies two square-root branches
For , these give and . Their derivatives follow by differentiating : , so . There are exactly two branches on : for any other holomorphic square root , the ratio is continuous, takes values in , and hence is constant on the connected domain.
The cut can move. Choosing defines a holomorphic logarithm on . At this gives and , both regular values despite lying on the previous cut. In the complex upper half-plane, agrees with ; in the lower half-plane it equals , so the corresponding square roots have opposite signs there. The overlap has two components, exactly why the identity theorem cannot force agreement between them. [ref]
For example, the limits of at , , are from above and from below. This jump records the chosen branch cut; each point still has a neighbourhood supporting a holomorphic square root. Moving the cut changes the bookkeeping, while the branch point at and its monodromy persist. Specifying a branch of a multivalued function is therefore essential before differentiating or performing contour integration with holomorphic square roots, complex logarithms or complex exponentiation.
Complex square root 2026-10-20
For a nonzero complex number with , its two square roots are . They are distinct and differ by a sign. At the only square root is zero. A principal value convention chooses and the plus sign; continuity across the negative real axis then fails. A holomorphic square root is a different, domain-dependent notion: it requires a holomorphic function whose square is on the specified domain. It cannot exist on every punctured neighborhood surrounding zero.
The monodromy theorem gives equality of continuations that exist under its hypothesis. It does not assert that every holomorphic function admits analytic continuation everywhere, even within a simply connected domain.
For an explicit obstruction, take near and attempt to apply the theorem on . The path ends at , where no terminal germ of a holomorphic function exists. If a holomorphic representative near continued , then the identity theorem would extend to that neighbourhood. Substituting would give . Thus the pole prevents the existence hypothesis, although is a simply connected domain. The analytic continuation article explains why the same rational function does continue throughout .
Existence at an endpoint by some route is also weaker than existence along every path from the base point. Along the real segment from towards , a holomorphic square root of tends to , yet cannot continue holomorphically through : a local representative would satisfy by the identity theorem, and differentiating at would give with . A finite limiting value therefore does not remove a branch point. More severe obstructions appear at a natural boundary of a holomorphic function.
Exhaust a simply connected hyperbolic Riemann surface in potential theory by compact bordered Jordan domains . Their Green functions on a Riemann surface have logarithmic poles and zero boundary values. Exponentiating their harmonic conjugates produces proper degree-one maps with . If , normalize . Domain monotonicity and a global positive Green function bound increasingly from above. Thus Montel theorem gives a nonconstant limit , which is injective by the locally uniform limit of univalent functions theorem. The Schwarz lemma makes extremal among normalized disk-valued maps. If omitted , taking a global holomorphic square root of a disk automorphism applied to , then renormalizing at , would multiply its derivative by . Thus is onto and gives the desired biholomorphism.
Holomorphic logarithm Created 2026-10-03 Updated 2026-10-18
A holomorphic logarithm of a nowhere-zero holomorphic function on a domain is a holomorphic function such that throughout . Differentiating with the chain rule gives . For , a holomorphic logarithm is a branch of the complex logarithm, and its derivative is .
The principal logarithm is holomorphic on the slit complex plane , with
Here is the chosen complex argument, and . One may conventionally assign a principal value on the negative real axis, but that axis is excluded from this holomorphic domain. [ref] For example, gives and ; all its other complex logarithms differ by , .
There is no holomorphic logarithm of on the punctured complex plane . Such a function would be a holomorphic primitive of , so its contour integral around every closed path would vanish. On the counterclockwise unit circle, however,
This is the period obstruction to a holomorphic antiderivative, and also exactly the logarithm change in the monodromy diagram.
Every nowhere-zero holomorphic function on a simply connected domain does admit a holomorphic logarithm. Choose and with , and define
The Cauchy integral theorem makes this contour integral independent of the path in . Then and ; its value at is , proving . Any two choices on a connected domain differ by a constant , since their difference is continuous and takes values in the discrete set . [ref]
For a computed example, take on the unit disc, with . Expanding as a geometric series and integrating gives
The construction also produces holomorphic square roots: satisfies , so here . More generally, gives a consistent branch of a multivalued function for complex exponentiation by . A holomorphic logarithm converts multiplicative expressions into additive ones while making the domain restrictions explicit.
Holomorphic square root Created 2026-10-03 Updated 2026-10-14
A holomorphic square root of a nonvanishing holomorphic function is a holomorphic function satisfying . It exists locally. On a connected open domain it exists globally exactly when the winding number of around zero is even for every closed path in the domain. Continuing a local square root once along such a path multiplies it by , where is that winding number. Thus the obstruction concerns parity, whereas a holomorphic logarithm needs every such winding number to be zero.
On , the function has no global holomorphic square root: its value winds once on the unit circle, and the lift shown under monodromy changes sign. In contrast, winds twice there and has global square roots and . It still has no holomorphic logarithm on that domain. This distinguishes the two global continuation conditions.
Holomorphic square root outside all polynomial roots Created 2026-10-07 Updated 2026-10-20
An even-degree polynomial with all roots in admits a holomorphic square root throughout . Factor out its even power of and use the convergent logarithm of each to construct it. Even degree removes the annular winding obstruction. The two choices differ by a global sign.
The Joukowski map restricts to a conformal map from onto the upper half-plane. Indeed, there. For any , the roots of have product one and cannot lie on the unit circle; exactly one has modulus greater than one, and that root has positive imaginary part. This proves that the map is a bijection.
Its inverse function is
with each factor taken as the principal square root. This product is the holomorphic square root of that behaves as at infinity in the upper half-plane. The upper unit semicircle maps to ; the exterior real rays map to the remaining real rays. The derivative vanishes only at the boundary endpoints .
Lemniscatic integral 2026-10-14
Choose the holomorphic square root equal to one at zero on the unit disc. The resulting elliptic integral maps that disc conformally onto the square with vertices , where . For the inverse Cayley transform between the half-plane and disk , the square of the derivative of , with , is . Thus the two square maps differ by an affine change of coordinate.
Start at with complex logarithm germ of a holomorphic function and holomorphic square root germ of a holomorphic function . On the punctured plane , both admit analytic continuation along every path. To construct it locally, for any choose a small open disc avoiding , select a holomorphic logarithm there, and select its additive multiple of to agree with the arriving germ of a holomorphic function; exponentiating half this logarithm supplies the arriving holomorphic square root. A finite chain of these open discs follows any given path.
Compare two paths from to , parametrized by :
The first follows a counterclockwise quarter-circle; the second follows a clockwise three-quarter-circle. Following and then path reversal of makes one positive turn about . Its winding number is , so the two paths cannot be homotopic with endpoints fixed in the punctured plane.
Figure 1.
Two paths from 1 to i enclose the puncture, while a negative-real-axis cut excludes the clockwise route and permits a deformation between the remaining routes
. Details.
The continued values along the chosen paths are
At their common endpoint,
Both complex logarithm values exponentiate to , and both holomorphic square root values square to . These give different terminal germ of a holomorphic function representatives, consistently obtained from the same starting germ of a holomorphic function. The monodromy theorem applies, but does not equate these two non-homotopic paths.
Figure 2.
Tracking continuation from 1 to i along a quarter-circle and a clockwise three-quarter-circle gives logarithms separated by 2 pi i and opposite square roots
. Details.
The closed path instead returns to , with continued values and . One counterclockwise turn changes the complex logarithm by and flips the holomorphic square root from to . Hence neither initial germ of a holomorphic function can come from a global holomorphic logarithm or a global holomorphic square root of on the punctured plane. Repeating a loop of winding number gives the changes and . [ref]
Now restrict to the slit domain . The principal logarithm and realize the same initial germ of a holomorphic function for each expression globally. To see that this is a simply connected domain, polar form identifies this domain with , a convex set; every loop contracts there. The simply connected corollary of the monodromy theorem now forces equality along every pair of paths with matching endpoints. In particular, and the segment both give and . The explicit homotopy stays in the closed first quadrant away from , hence in .
The other route crosses the removed ray at and is no longer admissible. In the punctured plane, that crossing is locally regular: analytic continuation passes through smoothly, although evaluating the principal logarithm would reset its argument. The branch cut changes the allowed domain; it does not create additional branch points or poles. That distinction is the beauty of the obstruction: perfectly compatible local holomorphic functions can fail to make a global choice solely because of the allowed based loops. The branches of a multivalued function article develops how moving the branch cut changes the selected values.
Monodromy Created 2026-10-03 Updated 2026-10-14
Monodromy records how a germ of a holomorphic function changes after analytic continuation around a closed path. The source point returns to itself, but the continued germ of a holomorphic function need not. For , , starting with holomorphic square root value gives and hence ; starting with complex logarithm value gives and hence . After counterclockwise turns the values are and . Thus monodromy exchanges the two square-root branches of a multivalued function, whereas it shifts through infinitely many logarithm branches of a multivalued function. [ref]
Figure 1.
One closed loop in the source plane lifts to an open semicircle for the square root and an open vertical segment for the logarithm
. Details.
The three panels use the same parameter . They show analytic continuation through local holomorphic patches, including across a chosen branch cut; repeatedly evaluating the principal logarithm would instead reset its imaginary part at the cut. Monodromy makes the obstruction to a global branch of a multivalued function visible and motivates the use of Riemann surfaces as domains on which continued values stay distinct.
It suffices to construct an injective holomorphic map for each domain. Then is a bounded entire function and is constant by Liouville's theorem; injectivity of forces to be constant.
For the upper half-plane, use the Cayley transform between the half-plane and disk
Its denominator is nonzero there, and , so its image lies in the unit disk. The determinant is nonzero, hence it is injective and conformal.
On the plane slit along the nonnegative real axis, choose the branch
This is the holomorphic square root obtained from the corresponding logarithm branch. Its argument lies in , so it maps bijectively onto the upper half-plane, with inverse . Its derivative is . Therefore
is an injective conformal map of onto the unit disk.
For the complement of the closed interval, first use
Its only finite pole is the excluded point one, and it is injective with derivative . If , solving gives , which is excluded. Hence , and
is the required injective conformal map into the disk. More precisely, and , since . No claim that the interval complement is conformally equivalent to the whole disk is required.
Applying Liouville to the three bounded compositions gives
This illustrates Liouville rigidity from a bounded holomorphic embedding: an explicit injective bounded coordinate is enough.
At large , , so the map tends to the identity. Its exterior inverse has the expansion . Therefore the physical complex potential of a two-dimensional potential flow is , with
Its leading term is unchanged, so the far-field velocity still has magnitude and inclination .
The chain rule gives the complex velocity:
The parameter denotes the unique exterior preimage of . If an expression solely in is desired, take
where the exterior holomorphic square root is chosen so that at infinity. Its branch points lie inside the ellipse. This is uniform potential flow around an ellipse.
The quadratic formula gives two algebraic possibilities. To select one consistent holomorphic square root, set
where each factor is the principal square root. Both arguments lie in the upper half-plane, so neither crosses its branch cut. This choice satisfies and at large in the upper half-plane. The required inverse function is
For , , this gives , whose modulus exceeds one. As varies continuously in the upper half-plane, this root cannot cross , since that would make real. Hence the selected root always lies in the fluid region. The other root is its reciprocal and lies inside the unit disk. Simply writing the principal square root of would not specify this branch correctly throughout the upper half-plane.
Use the Joukowski map with , in particular its intended range . On the original circle,
Define the exterior inverse
where the holomorphic square root is chosen so that at infinity. Then is the branch mapping the exterior of the ellipse to . Composition of the zero-circulation complex potential from part 4(c) with this inverse gives
The ellipse is a streamline because its inverse image is the circle, and at infinity. An additive constant in the complex potential has no effect on the velocity.
The transformed complex velocity obeys the chain rule:
Since , the denominator is nonzero on and outside . The four stagnation points therefore come from
Their physical locations are
They all lie on the ellipse, and their positions are independent of , which only sets the overall speed scale.
The printed restriction that is real and is insufficient literally: satisfies it but makes at in the exterior, so there is no exterior conformal map of the claimed type; the stated minor semiaxis would also be negative. The intended positive-parameter range, or more generally , is needed for the above solution. The reference to a sphere in the source is a wording slip: the geometry is two-dimensional flow around the mapped cylinder's elliptical cross-section.
Let , the inverse Cayley transform between the half-plane and disk. Direct calculation gives
so
On the unit disc, choose the holomorphic square root of equal to one at zero. Both sides define nonvanishing holomorphic derivatives, and their quotient has constant square. The branch chosen in part ii gives . Consequently
An affine change of image coordinate preserves conformality, so the lemniscatic integral maps the disc bijectively onto a square.
The normalization specifies the square exactly. Put . The substitution gives . The integral satisfies and , so its four corner values are . Thus
In particular the image is a square rotated through relative to the one in part ii, and its side length is . These symmetries also give from the affine identity.
Figure 1.
The lemniscate and the four-corner boundary correspondence of the lemniscatic disc-to-square map
. Details.
Choose the holomorphic square root in the complex upper half-plane whose boundary value is positive on . Its primitive has derivative . The integral from the boundary basepoint zero is well-defined by a limiting path: the local singularity is only .
The three finite prevertices have derivative exponent , hence image angle . At infinity , so the local coordinate also gives image angle . These are the four corners of a Schwarz-Christoffel mapping to a rectangle. To establish that it is a square rather than a general rectangle, put
The length along is also by . The length along is
by , and the fourth side has the same length by symmetry. With the chosen branch, the boundary derivatives on the intervals have phases , respectively. Therefore
The extended real boundary maps continuously and once around this square. Each side is traversed monotonically, and the four corner limits are finite. The argument principle, applied after small indentations around the boundary singularities, counts one inverse image of each interior value and zero of each exterior value. Since never vanishes in the complex upper half-plane, those interior inverse images are simple. Thus
The disc lies in the upper-right half-plane and avoids zero and the negative real axis. On the slit plane take the holomorphic logarithm with . Its derivative is , and . Therefore the holomorphic square root
is analytic and satisfies . One can also construct it without a global branch theorem: fix either scalar square root of and expand by its convergent binomial series. The series converges on the entire disc because . Either sign gives an admissible choice.
For a plane domain, the following are equivalent descriptions of a simply connected domain: every based loop contracts in the domain with its basepoint fixed; any two paths with the same endpoints are homotopic while those endpoints stay fixed; its complement in the Riemann sphere is connected; every holomorphic function has a single-valued primitive; and every nowhere-zero holomorphic function has a holomorphic logarithm. The complement must be taken in the extended plane, including infinity. An equivalent winding number formulation is that every closed curve has index zero about every point outside the domain.
Here is a direct equivalence proof for the first two topological formulations. If all paths with given endpoints are homotopic relative to those endpoints, apply this to a loop based at and the constant path at , obtaining a contraction. Conversely, let join to . The loop contracts with basepoint fixed. Concatenating this contraction with produces a homotopy from to , relative to . The middle retracing excursion contracts by shortening its traversal in both directions, giving a homotopy from that first path to . Hence and are homotopic relative to their endpoints. Planar domains are path-connected, so no extra path-component condition is needed.
The analytic descriptions follow from path independence of the holomorphic integral. Local primitives exist on discs by integrating power series; a homotopy can be subdivided into pieces lying in such discs, and local path integrals cancel along its common edges. Thus integrals over homotopic paths agree, giving a global primitive. Applied to for a nowhere-zero function, a primitive makes constant and hence supplies a holomorphic logarithm. These observations are also the tools used in the mapping proof below.
The Riemann mapping theorem states that every nonempty simply connected proper domain admits a conformal bijection onto the unit disc. Given , there is exactly one such map with and . The exclusion of the whole plane is necessary: an entire map into the disc would be constant by Liouville's theorem.
An extremal proof has the following concrete steps. First construct an injective bounded holomorphic map. Choose and a holomorphic square root of . It is injective, since equality of its values implies equality of their squares. Its image is disjoint from its negative: would imply and then , impossible. If , choose with ; the disc is omitted. Therefore is injective and bounded. Scaling it and applying an automorphism of the unit disk gives a map into the disc taking to zero; a rotation makes its derivative positive.
Let consist of all injective maps with this normalization. It is nonempty. The derivative at is bounded above by the Cauchy integral formula on a small disc compactly contained in . Choose with derivatives tending to their positive supremum . Boundedness, the Cauchy estimates and a diagonal Arzelà-Ascoli theorem argument give a subsequence converging locally uniformly to ; this is the bounded-family form of the Montel theorem. Then , , and the maximum modulus principle puts its image in the open disc. The limit is injective: if distinct had , apply Hurwitz's theorem near to , which has no zero there; the limit would vanish identically, contradicting .
Finally suppose is omitted by . It is nonzero since . Let and choose a holomorphic square root of . It is injective and takes values in the disc. Put , so , and normalize by a rotation. The new map's derivative magnitude at is
This square-root improvement of a normalized conformal map contradicts extremality. Thus is onto. Two normalized bijections differ by a disc automorphism fixing zero; Schwarz lemma applied to it and its inverse makes it a rotation, and positivity of the normalized derivatives forces that rotation to be the identity. This completes the requested proof outline.
We first derive the exterior area theorem for univalent functions directly from Green's theorem, then obtain the coefficient bound needed for the Koebe quarter theorem.
Suppose is univalent on and has a Laurent series
For , the image of , , is a regular simple closed curve. By the Jordan curve theorem the curve has one bounded and one unbounded complementary component. The image of lies in the unbounded component, since it is connected, avoids the curve by injectivity and contains points near infinity. It fills that component: any finite boundary limit of the image has preimages bounded by , and a convergent preimage subsequence must land on or else give an interior image point. Thus no boundary of the image lies inside that complementary component. The curve is positively oriented about its bounded interior, since the exterior is on the right of and a holomorphic map with nonzero derivative preserves orientation. Green's formula gives its enclosed area as
For completeness, the positive Fourier mode is and the negative modes are . On multiplying by the conjugate series and integrating, all unequal-frequency products integrate to zero by Fourier orthogonality. The equal-frequency products give and , respectively; the constant term gives zero. The Laurent series and its differentiated series converge uniformly on the circle, since , so these operations are justified. Nonnegativity of yields . Letting , or first doing so for every finite partial sum, gives
No regularity of the map on the unit circle has been assumed.
Now let be a normalized univalent function. Its only zero is at zero, so , extended there by the value one, is holomorphic and nowhere zero. On the simply connected disc it has a holomorphic square root with value one at zero. Its odd square-root transform is
This is an odd function which is holomorphic and normalized. It is also univalent: gives and hence . If , oddness gives , so , forcing . Thus is univalent on the exterior disc, with expansion
Applying the area bound to its first negative coefficient proves the second coefficient bound for normalized univalent functions, .
Let be any point omitted by . It is nonzero. Postcomposition with a Möbius map gives another normalized univalent function
whose denominator does not vanish. Applying the second coefficient bound both to and to gives
Every omitted point therefore has modulus at least , proving
This establishes the Koebe quarter theorem from the allowed area formula.
Sharpness is witnessed by the Koebe function . It has the required value and derivative at zero. The equality simplifies to , and makes it injective. Moreover,
Thus is omitted, so no larger universal radius is possible. In fact shows its image is the plane slit along .
Without univalence take a real and
It is entire, with and , but it omits , a point of modulus less than , since the exponential never vanishes. It is genuinely not univalent on the disc: the distinct points lie there and both give exponential value . Even its derivative never vanishes, illustrating why local conformality alone does not give the quarter-disc inclusion.