This chapter can be skipped if readers are only interested in geometry of two-dimensional surfaces in \(\R^3\).
The prefix hyper- in the word “hypersurface” means the manifold is one dimension lower than the ambient space. A hypersurface in \(\R^{n+1}\) is an \(n\)-dimensional subset of \(\R^{n+1}\). As in the case of regular surfaces in \(\R^3\), we want to impose some conditions on a hypersurface so that it becomes regular in the way we desire.
Let \(\Sigma^n\) be a non-empty subset of \(\R^{n+1}\). Suppose \(\Sigma^n\) can be covered by the image of a family of local parametrizations \(\mathcal{A}=\{F_\alpha:U_\alpha\to\Sigma\}\), i.e. \(\Sigma=\bigcup_\alpha F_\alpha(U_\alpha)\), such that each \(U_\alpha\) is an open set in \(\R^n\) and each \(F_\alpha:U_\alpha\to\Sigma\) satisfies all three conditions below:
(1) \(F_\alpha(u^1_\alpha,\cdots,u^n_\alpha)\) is a \(C^\infty\) map from \(U_\alpha\) to \(\R^{n+1}\);
(2) \(F_\alpha\) is a homeomorphism between \(U_\alpha\) and its image \(O_\alpha:=F_\alpha(U_\alpha)\); and
(3) The vectors \(\big\{\pd{F_\alpha}{u^1_\alpha},\cdots,\pd{F_\alpha}{u^n_\alpha}\big\}\) in \(\R^{n+1}\) are linearly independent on \(U_\alpha\).
Then, we say that \(\Sigma^n\) is a regular hypersurface in \(\R^{n+1}\).
Each vector \(\pd{F_\alpha}{u^i_\alpha}\big|_{(u^1_\alpha,\cdots,u^n_\alpha)}\) is a tangent to \(\Sigma^n\) at the based point \(p=F_\alpha(u^1_\alpha,\cdots,u^n_\alpha)\). With a slight abuse of notation and for simplicity, we will from now on denote \[\pd{F_\alpha}{u^i_\alpha}(p):=\pd{F_\alpha}{u^i_\alpha}\Big|_{(u^1_\alpha,\cdots,u^n_\alpha)}\] where \(p=F_\alpha(u^1_\alpha,\cdots,u^n_\alpha)\), to emphasize that \(p\) is the based point of the tangent vector.
By condition (3) in Definition 2.28, we know that the following vector space \[T_p\Sigma^n:=\Span\Big\{\pd{F_\alpha}{u^1_\alpha}(p),\cdots,\pd{F_\alpha}{u^n_\alpha}(p)\Big\}\] has dimension \(n\). This is called the tangent space at \(p\in\Sigma^n\).
The graph \(\Sigma_f\) of any smooth function \(f:\R^n\to\R\) is a regular hypersurface in \(\R^{n+1}\). One can parametrize \(\Sigma_f\) by a single parametrization: \[F:\R^n\to\Sigma_f,\qquad (x_1,\cdots,x_n)\mapsto\big(x_1,\cdots,x_n,f(x_1,\cdots,x_n)\big).\] By straightforward computations, we have \[\pd{F}{x_i}=\hat e_i+\pd{f}{x_i}\,\hat e_{n+1}\] where \(\{\hat e_1,\cdots,\hat e_{n+1}\}\) are the standard basis vectors in \(\R^{n+1}\). It is clear that \(\big\{\pd{F}{x_i}\big\}_{i=1}^{n}\) are linearly independent. \(\square\)
The \(n\)-dimensional unit sphere \[S^n:=\Big\{(x_1,\cdots,x_{n+1})\in\R^{n+1}:\textstyle\sum_{i=1}^{n+1}x_i^2=1\Big\}\] is a regular hypersurface in \(\R^{n+1}\). It can be parametrized by a pair of (inverse) stereographic projections \(F_+:\R^n\to S^n\setminus\{(0,\cdots,0,1)\}\) and \(F_-:\R^n\to S^n\setminus\{(0,\cdots,0,-1)\}\), with \(F_+\) given by \[F_+(u_1,\cdots,u_n)=\sum_{j=1}^{n}\frac{2u_j\,\hat e_j}{|u|^2+1}+\frac{|u|^2-1}{|u|^2+1}\,\hat e_{n+1}\] where \(|u|^2=u_1^2+\cdots+u_n^2\). We leave it as an exercise for readers to verify \(F_+\) satisfies the conditions in Definition 2.28 and write down the south-pole map \(F_-\). \(\square\)
Fill in the omitted detail in Examples 2.29 and 2.30.
Regular hypersurfaces are the higher-dimensional generalization of regular surfaces. As such many important results about regular surfaces can carry over naturally to regular hypersurfaces. We state the results below and leave the proofs as exercises.
Let \(g(x_1,\cdots,x_{n+1}):\R^{n+1}\to\R\) be a smooth function. Consider a non-empty level set \(g^{-1}(c)\) where \(c\) is a constant. If \(\grad g(p)\neq\mathbf 0\) at all points \(p\in g^{-1}(c)\), then the level set \(g^{-1}(c)\) is a regular hypersurface in \(\R^{n+1}\).
Assume all given conditions stated in Theorem 2.31. Furthermore, suppose \(F\) is a bijective map from an open set \(U\subset\R^n\) to an open set \(O\subset\Sigma:=g^{-1}(c)\) which satisfies conditions (1) and (3) in Definition 2.28. Then, \(F\) satisfies condition (2) as well and hence is a smooth local parametrization of \(g^{-1}(c)\).
Let \(\Sigma\subset\R^{n+1}\) be a regular hypersurface, and \(F_\alpha:U_\alpha\to M\) and \(F_\beta:U_\beta\to M\) be two smooth local parametrizations of \(\Sigma\) with overlapping images, i.e. \(W:=F_\alpha(U_\alpha)\cap F_\beta(U_\beta)\neq\varnothing\). Then, the transition maps defined below are also smooth maps: \[(F_\beta^{-1}\circ F_\alpha):F_\alpha^{-1}(W)\to F_\beta^{-1}(W),\qquad (F_\alpha^{-1}\circ F_\beta):F_\beta^{-1}(W)\to F_\alpha^{-1}(W).\]
Prove Theorem 2.31, Proposition 2.32, and Proposition 2.33.