Annals of Mathematics
A Complete Minimal Surface in R3 Between two Parallel Planes Author(s): Luquesio P. de M. Jorge and Frederico Xavier
Reviewed work(s):
Source: Annals of Mathematics, Second Series, Vol. 112, No. 1 (Jul., 1980), pp. 203-206 Published by: Annals of Mathematics
Stable URL: http://www.jstor.org/stable/1971325 . Accessed: 29/01/2013 09:41
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].
.
Annals of Mathematics is collaborating with JSTOR to digitize, preserve and extend access to Annals of Mathematics.
A complete
minimal
surface in R3
between two parallel planes
By LuQufEsio P. DE M. JORGE* and FREDERICO XAVIER**
E. Calabi has asked if it is possible to have a complete minimal surface in RX entirely contained in a half-space. We answer his question in the affir- mative. In fact, we prove considerably more, namely, we show how to produce complete minimal surfaces contained in slabs of R3. The proof was motivated by Remmert's ingenious idea of using Runge's theorem to exhibit proper analytic embeddings of the unit disc D c( C in C3, as explained in [11,
page 96.
It should be pointed out that M. Miranda 121 has shown that a surface that minimizes area globally and lies in a half-space of RX is a plane. Here, however, we deal with surfaces that minimize area only locally. We also observe that B. Lawson [51 has given examples of complete surfaces of con- stant mean curvature which lie between two parallel planes.
We take the oportunity to call attention to [4j where a related conjec- ture is discussed.
The Example
LEMMA 1. (Weierstrass representation of minimal surfaces). Let
f,
g: D SAC be holomorphic and set01
=-f(lg2) 02 = f(l + g2) 03= fg2 2
If f never vanishes then the function x(xl, X2, X3): D R3, where Xk Re
will define a (regular) minimal surface in R3 whose element of length is given by ds = X I dz I where
X If I + IgV12) 2
LEMMA 2. Let {DJ} be a sequence of closed discs centered at the origin, Dn ci Dn+1, U Dn = D. Let Kn ci D. be a compact set such that
Kn
nfDnl
= 5
0003-486X/80/0112-1/0203/004 $ 09.20/1
(? 1980 by Princeton University (Mathematics Department) For copying information, see inside back cover.
204 L. P. DE M. JORGE AND F. XAVIER
and D\K,, is connected. Let f be a holomorphic function defined in a neigh- borhood of the union of the K,. Then it is possible to approximate f uni- formly on that union by holomorphic functions defined in D.
For the proof of Lemma 1, see 3:1. Lemma 2 is precisely exercise G of [1], page 96; there the reader will also find a sketch of its proof using Runge's theorem.
THEOREM. There are nonfiat complete minimal surfaces of RK entirely
contained in a slab.
Proof. Consider the picture below.
Kn
As indicated, K,, is the compact region formed by deleting a piece from an annulus. Let
r,
be the difference between the outer and inner radii of the annulus. For the next compact set, that is, K,+,, we delete a piece from another annulus (disjoint from the first) but from the "other side", and so on. LetE =Uneven Kn and
0
= U,,oddK,,.
We shall say that a path a crosses K,, if it intercepts both the inner and outer arcs of the annulus from which K,, is obtained. A little reflection shows that
{KJ}
satisfies the following:(*) Any divergent path in D of finite Euclidean length will cross all but a finite number of the K,, in E or all but a finite number of the K,, in 0.
Note that if K, is chosen so that its "opening" is to the right of 0 then the segment
{(x, 0) ; - < x <
2
A COMPLETE MINIMAL SURFACE IN R3 205 Let {C"} be a sequence of positive numbers, to be specified later. By
Lemma 2 there is a holomorphic function h on D such that
I
h - c,[ I < 1 onK,,. Let eh be the function g appearing in Lemma 1 and set
f-
1/g. Con-sider the minimal surface determined by
f
and g, as in Lemma 1. An im- mediate consequence is that x3 is bounded; that is, our minimal surface M is contained in a slab. We will show that {Ca,} can be chosen so as to makeM complete. Let a be a divergent path in D, which may be supposed to have Euclidean speed one. We shall distinguish two cases.
1) Suppose that a has infinite Euclidean length, i.e., a: [0, oo) D. Since
1Ig + > 1,
2 \ g
it follows that
1 (a) = X(a(t)) dt =o
0
2) Suppose now that a has finite Euclidean length, i.e., a: [0, b)
D,
b < cO. Consider the first alternative given in * and let m be an integer such that a crosses every K& C E with n > m. From g = eh - ecne hcn wehave Ig
I
> eCn1 on K,,. LetJo = It e [0, b) I a(t) e K,}
-
Then
21(at) >- I g(at(t))|
d t
> En~,n vnS g(at(t)) Id t
?Enmnevenecnl dt >-n~ m,n even recnl
(recall that t is the Euclidean arc-length of a).
Similarly, if the other alternative in * holds we have
1(a) ? En!k,nodd r,,ecn1 .
Therefore, if cn is chosen to grow fast enough, the curve a will have infinite
length in M. In fact, it suffices to take c, = -log r-,,. UNIVERSIDADE FEDERAL DO CEARA, FORTALEZA-CE-BRAZIL UNIVERSIDADE FEDERAL DE PERNAMBUCO, RECIFE-PE-BRAZIL
REFERENCES
[ 1 ] K. HOFFMAN, Banach Spaces of Analytic Functions, Prentice-Hall, Inc., Englewood Cliffs,
N.J. (1962).
[21 M. MIRANDA, Frontiere minimali con ostacoli, Annali Dell UniversitA di Ferrara, Vol. XVI, N2 2 (1971), 29-37.
206 L. P. DE M. JORGE AND F. XAVIER
[ 4] L. P. M. JORGE and F. XAVIER, On the existence of complete bounded minimal surfaces
in RN, to appear in Bol. Soc. Bras. de Mat., Vol. 10, N' 2 (1979).