Generalization of the H.A. Schwarz Theorem on Stability of Minimal Surfaces

We proved two theorems on stability of minimal submanifolds in a Riemannian space, which can be included in a regular family of minimal submanifolds.

Збережено в:
Бібліографічні деталі
Дата:2007
Автори: Aminov, Yu., Witkowska, J.
Формат: Стаття
Мова:English
Опубліковано: Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України 2007
Онлайн доступ:http://dspace.nbuv.gov.ua/handle/123456789/7615
Теги: Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
Назва журналу:Digital Library of Periodicals of National Academy of Sciences of Ukraine
Цитувати:Generalization of the H. A. Schwarz theorem on stability of minimal surfaces / Yu. Aminov, J. Witkowska // Журн. мат. физики, анализа, геометрии. — 2007. — Т. 3, № 4. — С. 399-410. — Бібліогр.: 16 назв. — англ.

Репозитарії

Digital Library of Periodicals of National Academy of Sciences of Ukraine
id irk-123456789-7615
record_format dspace
spelling irk-123456789-76152010-04-07T12:01:04Z Generalization of the H.A. Schwarz Theorem on Stability of Minimal Surfaces Aminov, Yu. Witkowska, J. We proved two theorems on stability of minimal submanifolds in a Riemannian space, which can be included in a regular family of minimal submanifolds. Доведено дві теореми про стійкість мінімальних підбагатовидів в рімановому просторі, якщо мінімальний підбагатовид можливо включити в регулярну сім'ю мінімальних підбагатовидів. 2007 Article Generalization of the H. A. Schwarz theorem on stability of minimal surfaces / Yu. Aminov, J. Witkowska // Журн. мат. физики, анализа, геометрии. — 2007. — Т. 3, № 4. — С. 399-410. — Бібліогр.: 16 назв. — англ. 1812-9471 http://dspace.nbuv.gov.ua/handle/123456789/7615 en Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України
institution Digital Library of Periodicals of National Academy of Sciences of Ukraine
collection DSpace DC
language English
description We proved two theorems on stability of minimal submanifolds in a Riemannian space, which can be included in a regular family of minimal submanifolds.
format Article
author Aminov, Yu.
Witkowska, J.
spellingShingle Aminov, Yu.
Witkowska, J.
Generalization of the H.A. Schwarz Theorem on Stability of Minimal Surfaces
author_facet Aminov, Yu.
Witkowska, J.
author_sort Aminov, Yu.
title Generalization of the H.A. Schwarz Theorem on Stability of Minimal Surfaces
title_short Generalization of the H.A. Schwarz Theorem on Stability of Minimal Surfaces
title_full Generalization of the H.A. Schwarz Theorem on Stability of Minimal Surfaces
title_fullStr Generalization of the H.A. Schwarz Theorem on Stability of Minimal Surfaces
title_full_unstemmed Generalization of the H.A. Schwarz Theorem on Stability of Minimal Surfaces
title_sort generalization of the h.a. schwarz theorem on stability of minimal surfaces
publisher Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України
publishDate 2007
url http://dspace.nbuv.gov.ua/handle/123456789/7615
citation_txt Generalization of the H. A. Schwarz theorem on stability of minimal surfaces / Yu. Aminov, J. Witkowska // Журн. мат. физики, анализа, геометрии. — 2007. — Т. 3, № 4. — С. 399-410. — Бібліогр.: 16 назв. — англ.
work_keys_str_mv AT aminovyu generalizationofthehaschwarztheoremonstabilityofminimalsurfaces
AT witkowskaj generalizationofthehaschwarztheoremonstabilityofminimalsurfaces
first_indexed 2025-07-02T10:25:47Z
last_indexed 2025-07-02T10:25:47Z
_version_ 1836530477910982656
fulltext Journal of Mathematical Physics, Analysis, Geometry 2007, vol. 3, No. 4, pp. 399�410 Generalization of the H.A. Schwarz Theorem on Stability of Minimal Surfaces Yuriy Aminov Institute of Mathematics of the Polish Academy of Sciences 8 Sniadeckich, PO Box 21, 00-956, Warsaw, Poland E-mail:Y.Aminov@impan.gov.pl Mathematical Division, B. Verkin Institute for Low Temperature Physics and Engineering National Academy of Sciences of Ukraine 47 Lenin Ave., Kharkiv, 61103, Ukraine E-mail:aminov@ilt.kharkov.u Joanna Witkowska Bialystok University of Finance and Management, Faculty of Engineering 1 Grunwaldska Str., 19-300 Elk, Poland Received January 19, 2007 We proved two theorems on stability of minimal submanifolds in a Riemannian space, which can be included in a regular family of minimal submanifolds. Key words: Riemannian space, minimal submanifold, stability. Mathematics Subject Classi�cation 2000: 53A10, 53A07, 53C42. 1. Introduction H.A. Schwarz proved stability of a minimal surface in 3-dimensional Euclidean space E3, when this minimal surface could be included in a regular family of minimal surfaces [1]. It follows from this theorem that every compact domain on a minimal surface z = z(x1; x2) is stable. Notice, that the question of minimal surface stability was considered in [3�9]. The existence and applications of stable minimal surfaces were given in [10�16]. Here we give the generalizations of this theorem for the cases of minimal hy- persurfaces in a Riemannian space and for 2-dimensional surfaces in 4-dimensional Riemannian space. Let F n be a minimal submanifold with boundary � in a Riemannian manifold V N . We consider some submanifold �n with the same boundary �, which is close to F n in the class C1. c Yu. Aminov and J. Witkowska, 2007 Yu. Aminov and J. Witkowska We say that a compact domain D with nonempty boundary � on a minimal submanifold F n is stable, if for all submanifolds �n with the same boundary �, close to D in the class C1 but di�erent from D, the volume V ol(�n) is grater than the volume of D V ol(�n) > V ol(D): Theorem 1. If a simple connected compact domain D on an orientable minimal hypersurface F n in the Riemannian manifold V n+1 can be included in a regular family of minimal hypersurfaces, then this domain D is stable. Theorem 2. Let F 2 be an orientable minimal surface in 4-dimensional Riemannian manifold. Let a simple connected compact domain D on F 2 can be included in a 2-parametric regular family of minimal surfaces with integrable distribution of normal planes. Then this domain D is stable. Later we construct a 2-parametric family of stable minimal surfaces in Euclidean space E4 with nonintegrable distribution of normal planes. From another side, there exists a nonstable minimal surface in the Euclidean space E4, which can be included in the regular family of minimal surfaces. In this case the distribution of normal planes is nonintegrable, too. This example shows that the second condition in Th. 2 is essential. 2. Minimal Hypersurface Later under F n we understand the simple connected compact domain D. Let F n be included in the regular family of minimal hypersurfaces F n(t) such that F n(0) = F n. We introduce on F n a coordinate system with the coordi- nates y1; : : : ; yn. With the help of orthogonal trajectories to the family F n(t) we construct a coordinate system with the coordinates y1; : : : ; yn+1 in some neighbor- hood of F n. Every F n(t) corresponds to the equation yn+1 = const. The metric of the space V n+1 takes the following form: ds2 = nX i=1 aijdy idyj + an+1;n+1(dy n+1)2; (1) where all coe�cients depend on all coordinates as regular functions of the class C1. Later 1 � i; j � n. Denote an+1;n+1 = h; yn+1 = t. Lemma 1. The coe�cients Lij of the second quadratic form of F n(t) have the following form: Lij = � 1 2h @aij @t ; i; j = 1; : : : ; n: (2) 400 Journal of Mathematical Physics, Analysis, Geometry, 2007, vol. 3, No. 4 Generalization of the H.A. Schwarz Theorem on Stability of Minimal Surfaces This lemma is well known (see, for example, [2]). Denote a = jaij j. Lemma 2. If every F n(t) is a minimal hypersurface, then @a @t = 0: (3) For a minimal hypersurface the mean curvature H is equal to zero, and we have H = 1 n Lija ij = 0; where aij are the elements of the inverse matrix to jjaij jj. As a consequence of (2) and (3), we obtain aij @aij @t = 0: For simplicity we denote @aij @t = a 0 ij . Introduce also the following vectors: li = (a1i; : : : ; ani); l 0 i = (a 0 1i; : : : ; a 0 ni): Later we write these vectors in the form of columns and denote the determinant by [ ]. We have evidently @a @t = [l 0 1; l2; : : : ; ln] + � � �+ [l1; : : : ; l 0 n ] = a nX i;j aij @aij @t = 0: Let �n be some hypersurface, which is close to F n in the class C1. In this case �n has one-to-one projection on F n and in the correspondent points its tangent spaces are close. We can write the representation of �n in the evident form: yn+1 = f(y1; : : : ; yn) with the condition f j� = 0. Denote later @f @yi = fi. The �rst quadratic form dl2 = bijdy idyj of �n can be calculated with the help of metric form of V n+1 dl2 = nX i;j (aijdy idyj + h2fifjdy idyj): Hence bij = aij + h2fifj: Introduce the vectors ai = (a1i; : : : ; ani); m = (f1; : : : ; fn): Journal of Mathematical Physics, Analysis, Geometry, 2007, vol. 3, No. 4 401 Yu. Aminov and J. Witkowska Later these vectors are written in the form of columns. We have jbij j = [a1+h2f1m;a2+h2f2m; : : : ; an+h2fnm] = a+h2 nX i [a1; : : : ;m; : : : ; an]fi; where in sum the vector m stays on the i-th place. Taking the decomposition of every determinant in sum, we obtain jbijj = a(1 + h2 nX i;j=1 fifja ij): But the matrix aij is positively determined, so nX i;j=1 fifja ij � 0; (4) and the equality can be only in the case when all fi = 0. Therefore, f = const. But f j� = 0. Hence, f = 0. If we put a condition that �n is di�erent from F n, then there exists some subset, where in (4) we have strong inequality. Denote by G the domain of the coordinates y1; : : : ; yn. Now we can calculate the volume of �n and compare it with the volume of F n V ol(�n) = Z G q jbij(y1; : : : ; f)jdy 1 : : : dyn > Z G p a(y1; : : : ; f)dy1 : : : dyn = Z G p a(y1; : : : ; 0)dy1 : : : dyn = V ol(F n): Hence, F n is the stable minimal hypersurface. The reviewer remarked that in the paper by H. Rosenberg [17] there were some statements close to the ones of Th. 1. But in the paper there was indicated only a weak stability. Besides, the consideration was too short and therefore not clear enough. 3. Minimal Surface in a 4-Dimensional Riemannian Space Let F 2 be a minimal surface in the Riemannian 4-dimensional space V 4. We suppose that F 2 is included in a 2-parametric regular family of minimal surfaces F 2(t; �) in some neighborhood D such that F 2(0; 0) = F 2. We say that it is the �rst family. Through every point in the neighborhood D of F 2 there goes one and only one surface from this family. Therefore, at this point the normal plane is determined, and we have a distribution of normal planes. 402 Journal of Mathematical Physics, Analysis, Geometry, 2007, vol. 3, No. 4 Generalization of the H.A. Schwarz Theorem on Stability of Minimal Surfaces By the conditions of Th. 2, the distribution of these normal planes is integrable. So, there exists the second family of the surfaces which are orthogonal to the surfaces from the �rst family. With the help of these two families, in the same way we can construct the coordinates in the considered neighborhood. We take the coordinate system y1; y2 on the surface F 2 and take the surface N0 from the second family, which goes through some point p0 2 F 2. We introduce the coordinates y3; y4 on the surface N0. So, if a point p 2 D, then through this point p there goes one surface from the second family, which intersects with F 2 at the point with coordinates y1; y2 as well as one surface from the �rst family, which intersects with N0 at the point with coordinates y3; y4. Hence the point p has coordinates y1; : : : ; y4. Later the Latin indexes have the value 1 or 2, and the Greek ones 3 or 4, respectively. Then the �rst quadratic form of V 4 will be ds2 = 2X i;j=1 aijdy idyj + 4X �;�=3 a��dy �dy� ; where all coe�cients depend on y1; : : : ; y4. Now let some surface �2 be close to F 2 and have the same boundary. We can represent �2 in the following form y� = f�(y1; y2); � = 3; 4; and f� = 0 on the boundary. Denoting the metric of �2 by dl2 = bijdy idyj we obtain bij = aij + a��y � ;iy � ;j ; where y� ;i are the derivatives with respect to coordinate yi. Let F n(t; �) be a mi- nimal surface, which goes through the point with coordinates y1; y2; y3; y4 on the surface �2. Lemma 3. Determinant a of the �rst quadratic form of F 2(t; �) does not depend on y3 and y4 @a @y3 = @a @y4 = 0: Let �k = f�� k g; k = 1; 2, be an orthogonal basis of normal plane of F 2(t; �) and Lk ij be the coe�cients of the second quadratic forms of F 2(t; �) with respect to this basis. Following the de�nition of the second quadratic forms (see [2]), we have two equations for � = 3; 4 y�;ij + �����y � ;i y�:j = Lk ij� � k ; Journal of Mathematical Physics, Analysis, Geometry, 2007, vol. 3, No. 4 403 Yu. Aminov and J. Witkowska where ��� �� are the Christo�el symbols of the metric of V 4. Here y� ;ij are the second covariant derivatives of the function y� with respect to the metric of F 2(t; �). We notice that every surface of this kind has the representation y3 = const; y4 = const: Hence y� ;i = 0; y� ;ij = @2y� @yi@yj � �k ij y� ;k = 0; � = 3; 4; where �k ij are the Christo�el symbols of the metric of F n(t; �). Besides, yi ;j = 0. So we have ���ij = Lk ij� � k : From the expressions of the Christo�el symbols we obtain 1 2 a��( @a�i @yj + @a�j @yi � @aij @y� ) = Lk ij� � k : But a�i = 0. Therefore � 1 2 a�� @aij @y� = Lk ij �� k : For a minimal surface we have Lk ij aij = 0; k = 1; 2: Therefore, we have the system of equations @a @y3 a33 + @a @y4 a34 = 0; @a @y3 a34 + @a @y4 a44 = 0: From here Lemma 3 follows. Now we have jbij j = ����� a11 + a��y � ;1y � ;1 ; a12 + a �y ;1 y� ;2 a21 + a��y � ;2y � ;1 ; a22 + a �y ;2 y� ;2 ����� : Denote by y�ji = y � ;k aki and p�� = ����� y� ;1; y � ;1 y� ;2; y � ;2 ����� : 404 Journal of Mathematical Physics, Analysis, Geometry, 2007, vol. 3, No. 4 Generalization of the H.A. Schwarz Theorem on Stability of Minimal Surfaces Then the expression of jbij j can be transformed to the following one: jbij j = a(1 + a��(y �j2y�;2 + y�;1y �j1) + 1 2a a��a �p � p��): Denote by grady� the gradient of the function y� with respect to coordinates y1; y2 and the metric of F 2(t; �). So we obtain jbij j = a[1 + a��(grady �; grady�) + 1 2a (a33a44 � (a34) 2)(p34)2]: Here the brackets () at the second member in the right side denote the scalar product in the metric aijdy idyj at a point of F 2(t; �). It is clear that the third term is nonnegative. Let us denote A = (grady3)2; B = (grady3; grady4); C = (grady4)2: The second term in the expression of jbij j has the form a(Aa33 + 2Ba34 + Ca44): We have evidently AC �B2 � 0; a33a44 � (a34) 2 � 0: Under these conditions the expression T = Aa33 + 2Ba34 + Ca44 � 0. Hence jbij j � a. If there is an equality here, then p34 = 0. In this case there exist some functions �(y1; y2) and ��(�) such that y� = ��(�); � = 3; 4: Under this condition the expression T has the form T = jgrad�j2����a��: So, from T = 0 we conclude that y� = const; � = 3; 4: But �2 is di�erent from F 2. Therefore, we have some subset, where jbij j > a. But a depends neither on y3, nor on y4. Therefore V ol(�2) > V ol(F 2). Theorem 2 is proved. 4. One Example Now we construct a 2-parametric family of minimal surfaces in E4 with the nonintegrable distribution of normal planes. Journal of Mathematical Physics, Analysis, Geometry, 2007, vol. 3, No. 4 405 Yu. Aminov and J. Witkowska Denote by xk the coordinates in E4 and z1 = x1 + ix2, z2 = x3 + ix4: Consider the family of minimal surfaces in E4 , which are given as level surfaces of an analytical function of two complex variables f(z1; z2) = c1 + ic2: We have two real equations �1 = Ref(z1; z2) = c1; �2 = Imf(z1; z2) = c2: It is a well-known fact that this surface is minimal and it is a holomorphic curve in E4. Every compact domain is an absolutely minimized area. Normal plane is determined by the following vectors: Xi = grad�i; i = 1; 2: Then the condition of integrability of distribution of normal planes has the following form rX2 X1 �rX1 X2 = �1X1 + �2X2 (5) with some coe�cients �k. We take the particular example f = z1z2 + z21 + z22 : Then evidently we obtain �1 = x1x3 � x2x4 + x21 � x22 + x23 � x24; �2 = x2x3 + x1x4 + 2x1x2 + 2x3x4: Consequently, grad�1 = (x3 + 2x1;�x4 � 2x2; x1 + 2x3;�x2 � 2x4); grad�2 = (x4 + 2x2; x3 + 2x1; x2 + 2x4; x1 + 2x3): (6) For the simplicity of notation denote �1 = �; �2 = . By calculation we obtain the matrices of the second derivatives for the functions � and jj @2� @xi@xj jj = 0 BB@ 2; 0; 1; 0 0; �2; 0; �1 1; 0; 2; 0 0; �1; 0; �2 1 CCA ; jj @2 @xi@xj jj = 0 BB@ 0; 2; 0; 1 2; 0; 1; 0 0; 1; 0; 2 1; 0; 2; 0 1 CCA : (7) 406 Journal of Mathematical Physics, Analysis, Geometry, 2007, vol. 3, No. 4 Generalization of the H.A. Schwarz Theorem on Stability of Minimal Surfaces Let us denote the second derivatives of functions, for example, �, by �ij. Intro- duce the notation ri = X j (�ij j � ij�j): (8) With the help of (5) - (8) we obtain the system of equations for �i r1 = 10x2 + 8x4 = �1(x3 + 2x1) + �2(x4 + 2x2); r2 = �10x1 � 8x3 = �1(�x4 � 2x2) + �2(x3 + 2x1); r3 = 8x2 + 10x4 = �1(x1 + 2x3) + �2(x2 + 2x4); r4 = �8x1 � 10x3 = �1(�x2 � 2x4) + �2(x1 + 2x3): From the �rst two equations we have �1 = � 80(x2x3 � x1x4) (x1 + 2x3)2 + (x2 + 2x4)2 : From the last two equations we �nd �1 = � 80(x2x3 � x1x4) (2x1 + x3)2 + (2x2 + x4)2 : These expressions are di�erent, so the system does not have any solution. Hence, the distribution of normal planes is nonintegrable. 5. Minimal Surfaces in E4 with Nonparametric Representation Let the minimal surface F 2 � E4 be given in the form x3 = u(x1; x2); x4 = v(x1; x2): Later we denote derivatives in the form of ui; uij . The functions u and v of a minimal surface satisfy two di�erential equations (see, for example, [10]) u11(1 + u22 + v22)� 2u12(u1u2 + v1v2) + u22(1 + u21 + v21) = 0; v11(1 + u22 + v22)� 2v12(u1u2 + v1v2) + v22(1 + u21 + v21) = 0: (9) It is easy to construct the family of minimal surfaces F 2(c1; c2) x3 = u+ c1; x4 = v + c2; ci = const: Journal of Mathematical Physics, Analysis, Geometry, 2007, vol. 3, No. 4 407 Yu. Aminov and J. Witkowska A normal plane is determined by vectors X1; X2 X1 = (u1; u2;�1; 0); X2 = (v1; v2; 0;�1): The condition of integrability of the distribution of normal planes is represented by the following system of equations: u11v1 + u12v2 � v11u1 � v12u2 = �1u1 + �2v1; u12v1 + u22v2 � v12u1 � v22u2 = �1u2 + �2v2; 0 = ��1 + 0�2; (10) 0 = 0�1 � �2: From here we have �1 = �2 = 0. Hence the condition (10) has the form u11v1 + u12v2 = v11u1 + v12u2; u12v1 + u22v2 = v12u1 + v22u2: (11) Therefore, by Theorem 2 the minimal surface in E4 at nonparametric repre- sentation is strongly stable if it satis�es the system of equations (11). The reviewer proposed to construct an example of minimal surface which would satisfy the system (9),(11). To construct this example we put u = �(x1) + �(x2); v = �(x1) + �(x2): Then the system (9),(11) has the following form: �00(1 + �02 + �02) + �00(1 + �02 + �02) = 0; �00(1 + �02 + �02) + �00(1 + �02 + �02 = 0; �00�0 � �00�0 = 0; �00�0 � �00�0 = 0; where 0 (prime) denotes the derivatives of function � or �; : : : with respect to their arguments. From the third and forth equations we obtain � = C1� + C2; � = C3� + C4; 408 Journal of Mathematical Physics, Analysis, Geometry, 2007, vol. 3, No. 4 Generalization of the H.A. Schwarz Theorem on Stability of Minimal Surfaces where Ci are constants. After substitution � and � into the �rst and the second equations we conclude that C1 = C3 and the second equation can be rewritten in the form of equation with separate arguments �00 1 + �02a2 = � �00 1 + �02a2 = k; where k = const; a = p 1 +C2 1 . By integration we obtain the equation of minimal surface u = p a2 � 1v; v = 1 ka2 ln cos(kax2 + d2) cos(kax1 + d1) ; where di are constants. It is evidently that the surface is not determined on the whole plane x1, x2. In [7] M.J. Micallef proved the following Corollary 5.1 A complete stable minimal surface in E4, which is an entire graph, is holomorphic. He indicated that in [10] R. Osserman constructed the examples of entire two-dimensional minimal graphs in E4, which were not holomorphic with respect to any orthogonal complex structure on E4. These graphs are unstable by Cor. 5.1. So, on this surface there exist the unstable domains. One of the Osserman surfaces has the following representation: x3 = u = 1 2 cos x2 2 (ex1 � 3e�x1); x4 = v = � 1 2 sin x2 2 (ex1 � 3e�x1): It is possible to include this surface in the family of minimal surfaces. The dis- tribution of normal planes is not integrable, because the equations (11) for this surface are not satis�ed. Therefore, the condition of integrability of the distribu- tion of normal planes in Th. 2 is essential. The Authors are thankful to the reviewer for helpful remarks. References [1] H.A. Schwarz, Gesammelte Mathematische Abhandlungen. Springer, Berlin, 1890. [2] L.P. Eisenhart, Riemannian Geometry. Princeton Univ. Press, Princeton, NJ, 1949, . [3] J.L. Barbosa and M. do Carmo, On the Size of a Stable Minimal Surface in R3. �Amer. J. Math. 98(2) (1976), 515�528. [4] M. do Carmo and C.K. Peng, Stable Minimal Surfaces in R3 are Planes. � Bull. Amer. Math. Sos. 1 (1979), No. 6, 903�906. [5] A.V. Pogorelov, On Stability of Minimal Surfaces. � Dokl. AN USSR 260 (1981), No. 2, 293�295. (Russian) Journal of Mathematical Physics, Analysis, Geometry, 2007, vol. 3, No. 4 409 Yu. Aminov and J. Witkowska [6] M. do Carmo, Stability of Minimal Submanifolds. � Lect. Not. Math. 838 (1984), 129�139. [7] M. Micallef, Stable Minimal Surfaces in Euclidean Space. � J. Di�. Geom. 19 (1984), 57�84. [8] Yu.A. Aminov, Geometry of the Submanifolds. Gordon and Breach Acad. Publ. House, Amsterdam, 2001. [9] D. Fischer-Colbrie and R. Schoen, The Structure of Complete Stable Minimal Surfaces in 3-Manifolds of Nonnegative Scalar Curvature. � Comm. Pure Appl. Math. 33 (1980), 199�211. [10] R. Osserman, A Survey of Minimal Surfaces. Van Nostrand Reinhold, New York, 1969. [11] J.D. Moore, On Stability of Minimal Spheres and a Two-Dimensional Version of Synge's Theorem. � Archiv. der Mat. 44 (1985), 278�281. [12] J.D. Moore, Compact Riemannian Manifolds with Positive Curvature Operators. �Bull. AMS 14 (1986), 279�282. [13] R.M. Schoen and S.-T. Yau, Existence of Incompressible Minimal Surfaces and the Topology of Three Dimensional Manifolds with Nonnegative Scalar Curvature. �Ann. Math. 110 (1979), 127�142. [14] J. Sacks and K. Ulenbeck, The Existence of Minimal Immersions of 2-Spheres. �Ann. Math. 113 (1981), 1�24. [15] M. Micallef and J.D. Moore, Minimal Two-Spheres and the Topology of Manifolds with Positive Curvature on Totally Isotropic Two-Planes. � Ann. Math. 127 (1988), 199�227. [16] M.J. Micallef and J.G. Wolfson, The Second Variation of Area of Minimal Surfaces in Four-Manifolds. � Math. Ann. 295 (1993), 245�267. 410 Journal of Mathematical Physics, Analysis, Geometry, 2007, vol. 3, No. 4