On the area of constant mean curvature discs and annuli with circular boundaries (Q5947802)

A natural question to ask is whether a compact constant mean curvature (CMC) surface which is bounded by a circle is necessarily a spherical cap or a flat disc. A CMC surface with circular boundary is the mathematical model of a soap bubble which has its boundary on a round hoop, and the surfaces we almost always observe are spherical caps, so that it is natural to ask if these are the only solutions. In 1991 \textit{N. Kapouleas} [J. Differ. Geom. 33, 683-715 (1991; Zbl 0727.53063)] gave a negative answer to this question by showing that for each \(g>2\) there are infinitely many such surfaces of genus \(g\). However the original question remains open if one requires in addition that the surface has genus zero or that it is embedded. In the genus zero case \textit{R. López} and the authors recently showed that the only stable CMC surfaces of disc type which are bounded by a circle are spherical caps [Proc. Am. Math. Soc. 127, 1195-1200 (1999; Zbl 0938.53007)]. In the paper under review, the authors use the Faber-Krahn inequality to convert this stability result into a lower bound for the area of a non-spherical CMC disc type surface in the Riemannian space form \(M^3(c)\) which is bounded by a circle of radius \(r\). Specifically, if \(\Sigma\) is such a surface, they obtain (Theorem 2) that its area \(A(\Sigma)\) satisfies \[ A(\Sigma) \geq\max\bigl\{ A_c (r,H), \widetilde A_c(r,H) \bigr\}, \] where \(A_c(r,H):= {4\pi\over H^2+c}-{\pi \over H^2+c} \sqrt{1-r^2 (H^2+c)}\), and \(\widetilde A_c(r,H): ={2\pi\over H^2+c} +{2\pi \over H^2+c}\sqrt {1-r^2 (H^2+c)}\). A similar idea is used to give a lower bound for the area of a nonzero CMC annulus in the Euclidean space \(E^3\) which is bounded by circles of radii \(r_1,r_2\) which project to concentric circles in a plane. If \(\Sigma\) is such a surface which is not a surface of revolution, the authors obtain (Theorem 3) that its area \(A(\Sigma)\) satisfies \[ A(\Sigma)\geq B(r_1, r_2,H,P): ={2\pi\over H^2}-{\pi\over H^2}\sum_{j=1,2} \left(1- \left(Hr_j+ {P\over r_j}\right)^2 \right)^{1/2}, \] where \(P\) is a ``flux parameter''. As a consequence, they can show that for certain values \((r_1,r_2, H,P)\) Delaunay surfaces are the unique area minimizers with these given parameters.