Chapter 3 › 3.2 Second Fundamental Form and Curvatures › 3.2.2 Curvatures of Graphs
Subsection

3.2.2  Curvatures of Graphs

This subsection assumes \(\Sigma\) is a two-dimensional regular surface in \(\mathbb{R}^3\). In Examples 3.4 and 3.17, we computed the first and second fundamental forms of the graph \(\Sigma_f\) of a function \(f\). Using these, it is not difficult to compute various curvatures of the graph. In this subsection, we are going to discuss the geometric meaning of each curvature in this context, especially at the point where the tangent plane is horizontal.

Proposition 3.24

Let \(\Sigma_f\) be the graph of a two-variable function \(f(u_1,u_2): \mathcal{U}\subset\mathbb{R}^2\to\mathbb{R}\). Suppose \(p\) is a point on \(\Sigma_f\) such that the tangent plane \(\Tp{p}\Sigma_f\) is horizontal, i.e. \(p\) is a critical point of \(f\), and the Gauss map \(\nu\) is taken to be upward-pointing, then

  • \(K(p)>0\) and \(H(p)>0 \Rightarrow p\) is a local minimum of \(f\)
  • \(K(p)>0\) and \(H(p)<0 \Rightarrow p\) is a local maximum of \(f\)
  • \(K(p)<0 \Rightarrow p\) is a saddle of \(f\)

Checkpoint
Let \(p\) be a critical point of \(f\) (horizontal tangent plane) on the graph surface \(\Sigma_f\), with upward Gauss map. According to Proposition 3.24, which condition implies \(p\) is a saddle?
WhyBy Proposition 3.24, \(K(p)<0\) means the principal curvatures have opposite signs, so the surface curves upward in one direction and downward in another — exactly the signature of a saddle point of \(f\).
Proof — revealed step by step

At a critical point \(p\) of \(f\), we have \(\nabla f(p)=0\). From Examples 3.4 and 3.17: \[ g_{ij}(p) = \delta_{ij} + \underbrace{\pd{f}{u_i}(p)\pd{f}{u_j}(p)}_{=0} = \delta_{ij}, \qquad h_{ij}(p) = \frac{\pd{^2 f}{u_i\partial u_j}(p)}{\sqrt{1+|\nabla f(p)|^2}} = \pd{^2 f}{u_i\partial u_j}(p). \]

Therefore: \[ K(p) = \frac{\det[h]}{\det[g]}(p) = \det\left[\pd{^2 f}{u_i\partial u_j}(p)\right] = (f_{11}f_{22}-f_{12}^2)(p), \] \[ H(p) = g^{ij}(p)h_{ij}(p) = \delta_{ij}\pd{^2 f}{u_i\partial u_j}(p) = (f_{11}+f_{22})(p). \]

From the second derivative test in multivariable calculus: if \(f_{11}f_{22}-f_{12}^2>0\) and \(f_{11}+f_{22}>0\) at \(p\), then \(p\) is a local minimum. The other cases follow similarly. \(\square\)

Given any regular surface \(\Sigma\) (not necessarily the graph of a function) and any point \(p\in\Sigma\), one can apply a rigid-body motion \(\Phi:\mathbb{R}^3\to\mathbb{R}^3\) so that \(\Tp{p}\Sigma\) is transformed into a horizontal plane. Then, the new surface \(\Phi(\Sigma)\) becomes locally a graph of a function \(f\) near the point \(\Phi(p)\). Recall that the Gauss curvatures of \(p\) and \(\Phi(p)\) are the same as given by Exercise 3.15. If \(K(p)>0\) (and hence \(K(\Phi(p))>0\)), then Proposition 3.24 asserts that \(\Phi(p)\) is a local maximum or minimum of the function \(f\) and so the surface \(\Phi(\Sigma)\) is locally above or below the tangent plane at \(\Phi(p)\). In other words, near \(p\) the surface \(\Sigma\) is locally on one side of the tangent plane \(\Tp{p}\Sigma\). On the other hand, if \(K(p)<0\) then no matter how close to \(p\) the surface \(\Sigma\) would intersect \(\Tp{p}\Sigma\) at points other than \(p\).