Available online at www.sciencedirect.

com

Indagationes Mathematicae 24 (2013) 415–427

www.elsevier.com/locate/indag

The Gauss map of a harmonic surface

David Kalaj ∗
Faculty of Natural Sciences and Mathematics, University of Montenegro, Cetinjski put b.b. 81000 Podgorica,
Montenegro

Received 31 January 2012; received in revised form 21 December 2012; accepted 8 January 2013

Communicated by E.P. van den Ban

Abstract

We prove that the distortion function of the Gauss map of a surface parametrized by harmonic
coordinates coincides with the distortion function of the parametrization. Consequently, the Gauss map
of a harmonic surface is K quasiconformal if and only if its harmonic parametrization is K quasiconformal,
provided that the Gauss map is regular or what is shown to be the same, provided that the surface is non-
planar. This generalizes the classical result that the Gauss map of a minimal surface is a conformal mapping.
⃝c 2013 Royal Dutch Mathematical Society (KWG). Published by Elsevier B.V. All rights reserved.

Keywords: Gauss map; Quasiconformal harmonic surfaces; Minimal surfaces; Regular surfaces

1. Introduction and statement of the main result

Let (N , h) be a Riemannian manifold and (M, g) be a compact Riemannian manifold. A

smooth map X : M → N between  M and N is said to be harmonic if it is a critical point of
the Dirichlet functional E[X ] := M ∥d X ∥2 d Vg . This definition is extended to the case where
M is not compact by requiring the restriction of X to every compact domain to be harmonic, or,
more typically, requiring that X be a critical point of the energy functional in the Sobolev space
H21 (M, N ).
Equivalently, the map X is harmonic if it satisfies the Euler–Lagrange equations associated
to the functional E. Harmonic maps model the extrema of the energy functionals associated to
some physical phenomena in viscosity, dynamic of fluids, electromagnetism, cosmology.... When

∗ Tel.: +382 20264553.

E-mail addresses: davidkalaj@gmail.com, davidk@t-com.me.
c 2013 Royal Dutch Mathematical Society (KWG). Published by Elsevier B.V. All rights
0019-3577/$ - see front matter ⃝
reserved.
doi:10.1016/j.indag.2013.01.001
416 D. Kalaj / Indagationes Mathematicae 24 (2013) 415–427

N = Rn , the Euler–Lagrange equation is indeed the Dirichlet equation ∆X = 0, where ∆ is the

Laplace operator on M. See, for instance, the surveys [6] and the references therein for a good
setting.
If M has dimension two, the Dirichlet functional actually depends only on the conformal
structure of the surface. The interplay between conformality, minimality and harmonicity is
a core topic in geometrical analysis, which connects the complex analysis and differential
geometry. It is well known that a conformal immersion of a Riemann surface in the Euclidean
three-space is minimal if and only if it is harmonic, and in this case its Gauss map is conformal
(meromorphic). This paper is motivated by the recent paper of Alarcon and Lopez [2]. In that
paper the authors considered parametric harmonic quasiconformal immersions of surfaces in
the sense of Definition 1.1 into the Euclidean space R3 and prove some interesting facts. A
typical result from [2] is the following: A complete harmonic immersion X of an open Riemann
surface M into R3 has finite total curvature if and only if it is algebraic and quasiconformal.
In this paper we will consider the class of parametric harmonic quasiconformal surfaces in
the sense of the standard definition of quasiconformality. Our approach is completely different
from and complementary to the approach of Alarcon and Lopez. Among others, we will show
that the definition of quasiconformality of harmonic parametrization of Alarcon and Lopez
(Definition 1.1), coincides with the standard definition of quasiconformality of parametrization
provided that the surface is not planar. On the other hand this result can be considered as
an extension of a classical result for minimal surfaces: if the minimal surface is given by
Enneper–Weierstrass parametrization, then its complex Gauss map is a meromorphic function.

1.1. Parametric and harmonic surfaces

We define an oriented parametric surface M in R3 to be an equivalence class of mappings

X = (a, b, c) : Ω → R3 of some domain Ω ⊂ C into R3 , where the coordinate functions a, b,
c are of class at least C 1 (Ω ). Two such mappings X : Ω → R3 and X̃ : Ω̃ → R3 , referred to
as parametrizations of the surface, are said to be equivalent if there is a C 1 -diffeomorphism
onto
φ : Ω̃ −→ Ω of positive Jacobian determinant such that X̃ = X ◦ φ. Let us call such a
diffeomorphism φ a change of variables, or reparametrization of the surface. Furthermore, we
assume that the branch (critical) points of M are isolated. These are the points (x, y) ∈ Ω
at which the tangent vectors X x = ∂Y ∂Y
∂ x , X y = ∂ y are linearly dependent or equivalently
X x × X y = 0 where by × is denoted the standard vectorial product in the space R3 . Equivalently,
at the critical points the Jacobian matrix ∇ X (z) has rank at most 1. It has full rank 2 at the
regular points. A surface with no critical points is called an immersion or a regular surface. If a
surface M allows a parametrization X = (x1 , x2 , x3 ) : Ω → M satisfying the Laplace equation
∆X = (0, 0, 0), then the surface is called harmonic surface. We will simultaneously use the
terminology harmonic surface for the surface M and for its harmonic parametrization X . The
same surface M could admit several harmonic parametrizations. For example if Ω is the unit
disk U and if X : U → M is a parametrization and ϕ is a Möbius transformation of the unit
disk, then f ◦ ϕ is a harmonic parametrization. A harmonic immersion is a regular harmonic
parametrization

1.2. Standard definition of quasiconformality

An orientation preserving smooth mapping ϕ : Ω → Ω ′ , between two open domains

Ω , Ω ′ ⊂ C is called K (K ≥ 1) quasiconformal (QC for short) if the dilatation d f of the
 D. Kalaj / Indagationes Mathematicae 24 (2013) 415–427 417

Beltrami coefficient µ(z) := ϕz̄ /ϕz satisfies the inequality

K−1
dϕ := |µ(z)| ≤ k := ,
K+1
or what is the same if
 
1 1
∥∇ϕ∥2 ≤ K+ Jϕ ,
2 K
where ∥∇ϕ∥ is the Hilbert–Schmidt norm defined by
∥∇ϕ∥2 := |ϕz |2 + |ϕz̄ |2
and where
Jϕ = |ϕz |2 − |ϕz̄ |2
is the Jacobian of ϕ.
Note that in our context it is enough to assume that ϕ is smooth mapping; however, the
classical definition of quasiconformality assumes weaker conditions (cf. [1, pp. 3, 23–24]). Note
also that, in this definition we do not assume injectivity. A parametric surface Y : Ω → R3 is
called K quasiconformal if
 
1
2 2
∥Yx ∥ + ∥Y y ∥ ≤ K + ∥Yx × Y y ∥, z = x + i y ∈ Ω . (1.1)
K
If K = 1 then (1.1) is equivalent to the system of the equations

∥Yx ∥2 = ∥Y y ∥2 and Yx , Y y = 0,
 
(1.2)
which represent isothermal (conformal) coordinates of the surface M = Y (Ω ). If Y is harmonic
and satisfies the system (1.2) then M = Y (Ω ) is a generalized minimal surface.
For regular points of the parametrization Y of a surface we define the distortion function by
∥Yx ∥2 + ∥Y y ∥2
DY = .
2∥Yx × Y y ∥
If Y is an immersion and X = (u, v, w) are isothermal coordinates of a smooth surface M, and
if ϕ = X −1 ◦ Y , then we have
|ϕz |2 + |ϕz̄ |2
D Y = Dϕ = . (1.3)
|ϕz |2 − |ϕz̄ |2
If we denote by f : S 2 \ {(0, 0, 1)} → R2 ,
 
x1 x2
f (x1 , x2 , x3 ) = ,
1 − x3 1 − x3
the stereographic projection, then the orientation preserving map g := f ◦ n is said to be the
complex Gauss map of X .
Since f is a conformal mapping, then g is quasiconformal if and only if n is quasiconformal
with the same constant of quasiregularity. Moreover

|gz |2 + |gz̄ |2
Dn = Dg = . (1.4)
|gz |2 − |gz̄ |2
418 D. Kalaj / Indagationes Mathematicae 24 (2013) 415–427

1.3. The Gauss map of a surface

Let X : Ω → R3 be a regular parametrization of the smooth surface M = X (Ω ). Let S 2 be

the unit sphere in R3 . Then the orientation preserving Gauss map n : Ω → S 2 of M is defined
as follows
Xu × Xv
n(z) = .
∥X u × X v ∥

1.4. Enneper–Weierstrass parametrization of minimal surfaces

A smooth enough surface can be parametrized using a isothermal parametrization X =

(x1 , x2 , x3 ). Such a parametrization is minimal if the coordinate functions xk , k = 1, 2, 3 are
harmonic, i.e., φk (z) := ∂∂zxk , k = 1, 2, 3 are analytic. Here ∂/∂z is the complex partial derivative
defined as
∂ 1 ∂ 1 ∂
 
:= + , z = x + i y.
∂z 2 ∂x i ∂y
A minimal surface can therefore be defined by a triple of analytic functions such that

φ12 + φ22 + φ32 = 0. (1.5)

The real parametrization is then obtained as

xk = Real φk (z)dz, k = 1, 2, 3. (1.6)

For an analytic function f and a meromorphic function g, the triple of functions

f
φ1 (z) = (1 − g 2 ) (1.7)
2
if
φ2 (z) = (1 + g 2 ) (1.8)
2
φ3 (z) = f g (1.9)
are analytic as long as f has a zero of order ≥ 2m at every pole of g of order m.
If φ1 , φ2 and φ3 have no real periods and f has a zero of order exactly 2m at every pole of g of
order m, then the following formula gives a minimal surface in terms of the Enneper–Weierstrass
parametrization
  
f 2 if
X = Real (1 − g ); (1 + g ); f g dz.
2
(1.10)
2 2
It is well known that if the surface is minimal endowed with Enneper–Weierstrass
parametrization, then its complex Gauss map is a meromorphic function (this is well-known
as a theorem of Bonnet (1860)). The complex Gauss map is given by g(w) = −i/g(w). See for
example [5, Sec. 9.3] or [3] for the above facts.
We now make the alternative definition of quasiconformality of harmonic surfaces.

Definition 1.1 ([2]). A harmonic immersion X : Ω → R3 is said to be quasiconformal if its

orientation preserving Gauss map n : Ω → S 2 is quasiconformal (or equivalently, if g is
 D. Kalaj / Indagationes Mathematicae 24 (2013) 415–427 419

quasiconformal). In this case, X is said to be a harmonic QC parametrization of the harmonic

surface X (Ω ).
In this paper, we prove that two definitions of quasiconformality (the standard definition and
Definition 1.1) are equivalent, provided that the surface is not planar, extending in this way the
theorem of Bonnet for harmonic quasiconformal surfaces (Theorem 1.2 and Corollary 1.3). We
will say that the Gauss map of the surface is regular if the set of branch points of n, i.e. the set
of points z such that ∂∂nx (z) × ∂n
∂ y (z) = 0 has zero measure. We will prove in Theorem 1.5, that
this condition is equivalent to the condition that a part of the surface, and consequently the whole
surface belongs to a plane.

Theorem 1.2. The dilatation and distortion function of the Gauss map n of a harmonic
immersion X coincides with the dilatation and distortion function of the surface itself, provided
that the Gauss map is regular. In other words
Dn (z) = DX (z) (1.11)
and
dn (z) = d X (z). (1.12)

Corollary 1.3. A harmonic immersion X of a surface M is K quasiconformal if and only if its

Gauss map is K quasiconformal, provided that the Gauss map is regular.

Remark 1.4. Theorem 1.2 and Corollary 1.3 have local character and we can assume that the
domain of X is the unit disk U.
We also prove the following theorem

Theorem 1.5. The Gauss map of a harmonic surface is regular if and only if the surface is not
planar. If the Gauss map of a surface is not regular, then it must be a constant vector.

Remark 1.6. From Theorem 1.5 we find out that, in Theorem 1.2, instead of the condition “the
Gauss map is regular” we can simply say “the surface is non-planar”. When we say that the
Gauss map is regular we mean that, the cross product d∂x n × dy
∂
n is not identically zero in some
subset of nonzero measure of the domain (and consequently in the whole domain, see the proof
of Theorem 1.5). However, if we take the surface
 2 
x3

1
X = x, − + x + y ,1 − x + y + y ,
2 2
3 2
then the normal n of X satisfies
∂ ∂
n× n = 0, (1.13)
dx dy
for y = −1/2. Thus, the branch points of the Gauss normal of harmonic surfaces are not isolated,
as in the case of minimal surfaces.

Remark 1.7. The class of minimal surfaces with Enneper–Weierstrass parametrization is a

special case of the class of quasiconformal harmonic surfaces. Namely in this case K = 1,
because the coordinates are isothermal (conformal). In this case the condition (1.13) reduces to
420 D. Kalaj / Indagationes Mathematicae 24 (2013) 415–427

the condition that g(w) = −i/g(w) = const, i.e. the Gauss normal is a constant. This implies
that the minimal surface is planar.
The class of quasiconformal harmonic mappings between complex domains and two-
dimensional surfaces has been a very active area of research in recent years. For some regularity
results for this class we refer to the following papers [7,8,14,13]. For some regularity results for
the class of minimal surfaces we refer to the papers of J. C. C. Nitsche [11,10].
Recall that by a definition of Alarcon and Lopez [2] a harmonic immersion X : Ω → R3 is
said to be quasiconformal if its orientation preserving Gauss map n : Ω → S 2 is quasiconformal
(or equivalently, if g is quasiconformal). Among other special features, quasiconformal harmonic
immersions are quasiminimal in the sense of Osserman [12]. Let w a harmonic diffeomorphism
of the unit disk onto itself which is not quasiconformal. Let X (z) = (Re(w), Im(w), 0). Then
n = (0, 0, 1) and therefore it is a 1-quasiconformal mapping. This mean that the fact that the
condition “the Gauss map is regular” is important in Theorem 1.2. In other words, two above
definitions of quasiconformality are equivalent, provided that the Gauss map is regular (up to
branch points) or what is the same, if the surface is not planar.

2. Proofs

Remark 2.1. As ⟨n, nx ⟩ = n, n y = 0, it follows that nx , n y are vectors belonging to the linear
 

spam of the vectors X x and X y . It follows from the Weingarten equation (in matrix form) that
 
a b
(nx , n y ) = (X x , X y )
c d
where
   
a b 1 M F − LG N F − MG
= .
c d EG − F2 L F − M E MF − NE
Here
E = ⟨X x , X x ⟩ , G = X y, X y , F = Xx, X y
   

and
L = ⟨X x x , n⟩ , N = X yy , n , M = Xxy, n .
   

As the surface is harmonic we have L + N = 0, and one may ask if this guarantees that the
matrix
 
a b
c d
is conformal, proving in this way the equality Dn = D X at one stroke. Unfortunately the given
matrix is not conformal in general for harmonic surfaces.
Proof of Theorem 1.2. Let
X (x, y) = (a(x, y), b(x, y), c(x, y)).
Without loss of generality we can assume that a(z) = x, z = (x, y). Namely, let φ be a analytic
mapping of the unit disk U into C such that
φ(z) = a(z) + i ã(z).
 D. Kalaj / Indagationes Mathematicae 24 (2013) 415–427 421

Here ã(z) is the harmonic conjugate of a(z). Take instead of

X (z) = (a(z), b(z), c(z))
X̃ = X ◦ φ −1 (z)
in some simply connected neighborhood Ω of some nonsingular point z of φ ′ (z). Then the first
coordinate of X̃ is x in Ω . Notice also that D X (z) = D X̃ (z).
Let
Xx × X y
n(z) = .
∥X x × X y ∥
Then n(z) is given by

 c y bx − b y cx cy

 2  2 , −   2 ,
b y + c2y + c y bx − b y cx b2y + c2y + c y bx − b y cx

by 
  2  .
b2y + c2y + c y bx − b y cx
Define
d d
P= n(z), Q= n(z).
dx dy
The distortion function is
∥P∥2 + ∥Q∥2
Dn = .
∥P × Q∥
Then
N
∥P∥2 + ∥Q∥2 = (2.1)
G4
and
M
∥P × Q∥ = (2.2)
G3
where
  
c2y b2x y − b yy bx x + b y c y −2bx y cx y + c yy bx x + b yy cx x
 
M =
 
+ b2y c2x y − c yy cx x ,

2
G = b y 2 + c y 2 + c y bx − b y cx


and
 
N = c4y b2x y + b2x x
    
+ b2y c2yy 1 + b2x + c2x − 2b y c yy bx cx y + 1 + b2y + b2x + c2x c2x y

− 2b y bx cx y cx x + b2y c2x x
422 D. Kalaj / Indagationes Mathematicae 24 (2013) 415–427

− 2c3y b yy cx bx y + cx bx y bx x + b y bx y cx y + bx x cx x
  
     
− 2b y c y 1 + b2x + c2x bx y cx y + b yy c yy 1 + b2x + c2x − b y bx cx y

+ b2y bx y cx y + bx x cx x − b y c yy bx bx y − cx cx y
    

 
− cx cx y cx x + bx cx y bx x + bx y cx x
    
+ c y b2yy 1 + b2x + c2x + 1 + b2x + c2x b2x y + 2b y b yy −bx bx y + cx cx y
2
 

 
+ b2y b2x y + c2x y + b2x x + c2x x + 2b y c yy cx bx y


− bx bx y bx x + cx cx y bx x + bx y cx x .
 

Let
A = b2x y − b yy bx x , B = bx y cx y + bx x cx x ,
 
C = c2x y − c yy cx x δ = 1 + b2x + c2x .

Then, since ∆b = ∆c = 0 we obtain

 
M = c2y A − 2b y c y B + b2y C

and
  
N = c4y A + b2y c2yy δ − 2b y c yy bx cx y + 1 + b2y + b2x + c2x c2x y

− 2b y bx cx y cx x + b2y c2x x
− 2c3y b yy cx bx y + cx bx y bx x + b y B
 
 
− 2b y c y (δ + b2y )B − Q

+ c y δ A + 2b y b yy −bx bx y + cx cx y
2
 


+ b2y (A + C) + 2b y c yy cx bx y − bx bx y bx x + cx cx y bx x + bx y cx x ,
  

where
Q = b y bx b yy cx y − b y c yy bx bx y − cx cx y − cx cx y cx x + bx cx y bx x + bx y cx x .
    

Further, by using again ∆b = ∆c = 0 we obtain

  
N = c4y A + b2y c2yy δ − 2b y c yy bx cx y + 1 + b2y + b2x + c2x c2x y

− 2b y bx cx y cx x + b2y c2x x − 2c3y b y B
 
− 2b y c y (δ + b2y )B − b y bx b yy cx y − Q
 D. Kalaj / Indagationes Mathematicae 24 (2013) 415–427 423

+ c2y δ A + 2b y b yy cx cx y + b2y (A + C)
 


  
+ 2b y c yy cx bx y + cx cx y bx x + bx y cx x

and consequently
   
N = c4y A + b2y c2yy δ + 1 + b2y + b2x + c2x c2x y + b2y c2x x − 2c3y b y B
 
− 2b y c y (δ + b y )B − b y bx b yy cx y − Q
2


+ c y δ A + 2b y b yy cx cx y + b2y (A + C)
2
 


  
+ 2b y c yy cx bx y + cx cx y bx x + bx y cx x

and
   
N = c4y A + b2y c2yy δ + 1 + b2y + b2x + c2x c2x y + b2y c2x x − 2c3y b y B
   
− 2b y c y (δ + b y )B + c y δ A + b y (A + C)
2 2 2

   
= c y A + b y (b y + δ)C − 2c y b y B − 2b y c y (δ + b y )B
4 2 2 3 2

 
+ c y δ A + b y (A + C)
2 2

= c2y (c2y + δ + b2y )A + b2y (b2y + δ + c2y )C − 2b y c y (δ + b2y + c2y )C

 
= γ c2y A + b2y C − 2b y c y C
where
γ = 1 + b2x + b2y + c2x + c2y .
As
N
Dn =
GM
we obtain finally
 
2 2
γ c y A + b y C − 2b y c y C
Dn =  
G c2 A + b2 C − 2b c C
y y y y

1 + b2y + c2y + b2x + c2x

=   2 = DX .
b2y + c2y + c y bx − b y cx
The last part of the theorem follows from the formulas (1.3) and (1.4). 

Corollary 2.2 (Bernstein Theorem for Nonparametric Harmonic Surfaces). If the harmonic
nonparametric surface over R2 is quasiconformal, then the surface is planar.
424 D. Kalaj / Indagationes Mathematicae 24 (2013) 415–427

Proof. From the previous theorem we find out that the Gauss map is quasiconformal. Then by a
theorem of L. Simon [15], X must be planar. 

Corollary 2.3. If X = (a, b, c) is a harmonic surface then

|∇a|2 + |∇b|2 + |∇c|2
D n =  2  2  2 .
2 b y ax − a y bx + −c y ax + a y cx + c y bx − b y cx

Proof of Theorem 1.5. As in the proof of Theorem 1.2, we can assume that the first coordinate
of the surface is x. Further, by using the same notation
A = b2x y − b yy bx x , B = bx y cx y + bx x cx x , C = c2x y − c yy cx x

2
G = b y 2 + c y 2 + c y bx − b y cx ,


and
 
M = c2y A − 2b y c y B + b2y C ,

we obtain that
∂ ∂
n× n=0
dx dy
if and only if M = 0. The last is equivalent with
√ √ √
(|c y | A − |b y | C)2 + 2(|b y ||c y | AC − b y c y B) = 0.
It is an elementary application of the Cauchy–Schwarz inequality that
√
AC ≥ B.
Moreover
AC = B 2
if and only if
bx x cx x = bx y cx y .
Thus M = 0 if and only if
√ √
(|c y | A − |b y | C) = 0
and
√
|b y ||c y | AC − b y c y B = 0.
Thus
b y ≥ 0, c y ≥ 0,
bx x cx x = bx y cx y
and
√ √
cy A − b y C = 0.
 D. Kalaj / Indagationes Mathematicae 24 (2013) 415–427 425

The cases b y = 0 and c y = 0 are trivial. So assume that b y ̸= 0 and c y ̸= 0. Combining the last
two equalities we arrive at the equality
c y b yy = b y c yy .
It follows that
d by
=0
dy c y
i.e.
by
= λ(x),
cy
for some real function λ(x) depending only on x. Since b y and c y are real analytic, it follows
that λ is a real analytic function, i.e.
∞

λ(x) = λn x n .
n=0
Furthermore
∞

cy = (an z n + bn z̄ n )
n=0
and therefore
∞
 ∞

b y (z) = λn x n (an z n + bn z̄ n ).
n=0 n=0
Let us show that λn = 0 for n ≥ 2.
We will use the following well-known fact. Any harmonic function has (locally) a unique
representation as a sum of homogeneous harmonic polynomials αn z n + βn z n . Since λ1 x(a1 z +
b1 z̄) is the only possible harmonic polynomial of degree 2 in the expression for b y , it must be
a1 z + b1 z̄ = 2a1 y. Further
n−1

λn−k x n−k (ak z k + bk z̄ k )
k=0
is a harmonic polynomial of degree n and is therefore equal to
αn z n + βn z̄ n .
Thus
n−1

λn−k (z + z̄)n−k (ak′ z k + bk′ z̄ k ) = αn z n + βn z̄ n .
k=0
Since the representation

qi j z i z̄ j
i, j

is unique, it follows that for n ≥ 2, λn = 0. Hence

by
= λ1 x + λ0 .
cy
426 D. Kalaj / Indagationes Mathematicae 24 (2013) 415–427

By a similar argument we find that, if λ1 ̸= 0, then c y = ωy + ν. In this case

ω 2
c(x, y) = (y − x 2 ) + νy + ν1 x + ν0
2
and
b y = (λ1 x + λ0 )(ωy + ν).
Thus
ω ω
b(x, y) = (λ1 x + λ0 )(y + ν/ω)2 − 2 (λ1 x + λ0 )3 .
2 6λ1
However b and c obtained in this case do not satisfy the equation M ≡ 0 in an open set. Namely
M = λ41 (ν + ωy)8 ̸≡ 0.
It remains to consider the case λ1 = 0. Then
b(z) = λ0 c(z) + ν(x),
implying that ν(x) = ν0 + ν1 x. Thus
X (z) = (x, λ0 c(z) + ν0 + ν1 x, c(z)).
Thus the Gauss map of the surface X is
 
 ν 1 1 
1
 , − ,
2 + ν12 2 + ν12 2 + ν12
 

implying that the surface is contained (locally) in some plane π . Let O be an orthogonal
transformation of R3 mapping the plane π onto the plane z = 0, and take Y = O ◦ X . The third
coordinate function Y3 of Y is a harmonic function, vanishing in an open set, and consequently
it vanishes in the whole domain of X . This implies that the whole surface is planar. 

2.1. An open problem

The classical Bernstein theorem for minimal surfaces asserts that, a nonparametric minimal
surface over the complex plane is a plane (see e.g. [4, p 13]). This theorem has been generalized in
various directions. One of the interesting generalizations concerns p-minimal surfaces (see [16]).
Since the hyperboloid X (x, y) = (x, y, x 2 − y 2 ) is a harmonic immersed surface over C, the
Bernstein theorem is not true for harmonic surfaces. If X : C → M ⊂ R3 is a conformal
parametrization of a “complete embedded” minimal surface, then M is either a plane or a
helicoid (this is one of the main theorems in the modern theory of minimal surfaces [9]). The
following question arises. Let X : C → R3 be a complete quasiconformal harmonic embedding,
must X (C) be either a plane or a helicoid?

Acknowledgment

I thank the referee for his comments and suggestions that helped to improve this paper.
 D. Kalaj / Indagationes Mathematicae 24 (2013) 415–427 427

References

[1] L.V. Ahlfors, Lectures on quasiconformal mappings, Manuscript prepared with the assistance of Clifford J. Earle,
Jr. Van Nostrand Mathematical Studies, No. 10, D. Van Nostrand Co., Inc., Toronto, Ont.-New York, London, 1966.
[2] A. Alarcon, F.J. Lopez, On harmonic quasiconformal immersions of surfaces in R3 , Trans. Amer. Math. Soc. 365
(2013) 1711–1742.
[3] D. Bshouty D, A. Weitsman, On the Gauss Map of Minimal Graphs, Complex Var. Elliptic Equ. 48 (4) (2003)
339–346. (8).
[4] T.H. Colding, W. Minicozzi, Minimal surfaces, in: Courant Lecture Notes in Mathematics., vol. 4, Courant Institute
of Mathematical Sciences, New York, NY, 1999, p. 124. viii.
[5] P. Duren, Harmonic Mappings in the Plane, Cambridge University Press, 2004.
[6] F. Hélein, J.C. Wood, Harmonic maps, in: Handbook of Global Analysis, 1213, Elsevier Sci. B. V, Amsterdam,
2008, pp. 417–491.
[7] D. Kalaj, Quasiconformal harmonic mapping between Jordan domains, Math. Z. 260 (2) (2008) 237–252.
[8] D. Kalaj, M. Mateljević, Inner estimate and quasiconformal harmonic maps between smooth domains, J. d’Analise
Math. 100 (2006) 117–132.
[9] W.H. Meeks III, H. Rosenberg, The uniqueness of the helicoid, Ann. of Math. 161 (2) (2005) 727–758.
[10] J.C.C. Nitsche, On new results in the theory of minimal surfaces, Bull. Amer. Math. Soc. 71 (1965) 195–270.
[11] J.C.C. Nitsche, The boundary behavior of minimal surfaces. Kellogg’s theorem and Branch points on the boundary,
Invent. Math. 8 (1969) 313–333.
[12] R. Osserman, on complete minimal surfaces, Arch. Ration. Mech. Anal. 13 (1963) 392–404.
[13] D. Partyka, K. Sakan, On bi-Lipschitz type inequalities for quasiconformal harmonic mappings, Ann. Acad. Sci.
Fenn. Math. 32 (2) (2007) 579–594.
[14] M. Pavlović, Boundary correspondence under harmonic quasiconformal homeomorfisms of the unit disc, Ann.
Acad. Sci. Fenn. Math. 27 (2002) 365–372.
[15] L. Simon, A Hölder estimate for quasiconformal maps between surfaces in Euclidean space, Acta Math. 139 (1–2)
(1977) 19–51.
[16] V.G. Tkachev, External geometry of p-minimal surfaces, in: Geometry from the Pacific Rim, Walter de Gruyter,
Berlin, 1997, pp. 363–375.