Chapter 3 › 3.1 Fundamental Forms › 3.1.3 Change of basis
Subsection

3.1.3  Change of basis

Recall that there could be several different local parametrizations of a given surface. Take the \(z=0\) plane as an example, one can parametrize by \(F(x,y)=(x,y,0)\), or using the polar coordinates \(G(r,\theta)=(r\cos\theta, r\sin\theta, 0)\). The tangent vectors \(\{\pdF{x}, \pdF{y}\}\) and \(\{\pd{G}{r}, \pd{G}{\theta}\}\) can be written in terms of each other using the chain rule: \[ \pd{G}{r} = \pd{\mathbf{F}}{x}\frac{\partial x}{\partial r} + \pd{\mathbf{F}}{y}\frac{\partial y}{\partial r} = \cos\theta\,\pdF{x} + \sin\theta\,\pdF{y} = \frac{x}{\sqrt{x^2+y^2}}\pdF{x} + \frac{y}{\sqrt{x^2+y^2}}\pdF{y}. \]

Exercise 3.5

Express \(\pd{G}{\theta}\) as a linear combination of \(\pdF{x}\) and \(\pdF{y}\). Answer: \(-y\pdF{x} + x\pdF{y}\). Please show the steps.

Need help?

The cotangent vectors \(\{dx, dy\}\) and \(\{dr, d\theta\}\) can also be written in terms of each other. Generally speaking, if a cotangent vector \(T\) outputs \(T(\pdF{x}) = a\) and \(T(\pdF{y}) = b\), then \(T = a\,dx + b\,dy\). Therefore, to express \(dr\) and \(d\theta\) in terms of \(\{dx,dy\}\), we compute how they act on \(\{\pdF{x}, \pdF{y}\}\) using the chain rule: \[ dr\!\left(\pdF{x}\right) = \frac{\partial r}{\partial x}, \qquad dr\!\left(\pdF{y}\right) = \frac{\partial r}{\partial y}. \] Therefore, \[ dr = \frac{\partial r}{\partial x}\,dx + \frac{\partial r}{\partial y}\,dy, \] which is the chain rule of differentials.

Exercise 3.6

Show that \(d\theta = \dfrac{\partial\theta}{\partial x}\,dx + \dfrac{\partial\theta}{\partial y}\,dy\).

Need help?

Consider the plane \(z=0\), parametrized in two ways by \(F(x,y)\) and \(G(r,\theta)\). In \((x,y)\) coordinates the first fundamental form in tensor notations is \(dx\otimes dx + dy\otimes dy\). In \((r,\theta)\)-coordinates it is \(dr\otimes dr + r^2\,d\theta\otimes d\theta\). Using \(r=\sqrt{x^2+y^2}\) and \(\tan\theta = y/x\) we have \(\frac{\partial r}{\partial x}=\frac{x}{r}\), \(\frac{\partial r}{\partial y}=\frac{y}{r}\), \(\frac{\partial\theta}{\partial x}=\frac{-y}{r^2}\), \(\frac{\partial\theta}{\partial y}=\frac{x}{r^2}\), and one computes: \[ dr\otimes dr + r^2\,d\theta\otimes d\theta = dx\otimes dx + dy\otimes dy. \] Therefore, the first fundamental form in tensor notations (at least for the plane \(z=0\)) is independent of local parametrizations, even though it is not the case in matrix notations.

In fact, such a “coincidence” is generally true for any regular hypersurface \(\Sigma^n\) in \(\mathbb{R}^{n+1}\). Suppose \(F(x_1,\ldots,x_n)\) and \(G(u_1,\ldots,u_n)\) are two local parametrizations of \(\Sigma\). By the chain rule, \(\pdG{u_j} = \frac{\partial x_k}{\partial u_j}\pdF{x_k}\), and so \[ \left\langle\pd{G}{u_i},\pd{G}{u_j}\right\rangle = \frac{\partial x_k}{\partial u_i}\frac{\partial x_l}{\partial u_j}\left\langle\pdF{x_k},\pdF{x_l}\right\rangle. \] Also, by Exercise 3.7, \(du^i = \frac{\partial u_i}{\partial x_p}\,dx^p\), hence \(du^i\otimes du^j = \frac{\partial u_i}{\partial x_p}\frac{\partial u_j}{\partial x_q}\,dx^p\otimes dx^q\). Combining and using \(\sum_i \frac{\partial x_k}{\partial u_i}\frac{\partial u_i}{\partial x_p} = \delta_{kp}\), we get \[ \left\langle\pd{G}{u_i},\pd{G}{u_j}\right\rangle du^i\otimes du^j = \left\langle\pdF{x_k},\pdF{x_l}\right\rangle dx^k\otimes dx^l. \] To conclude, even though the matrix representation \([g]\) depends on the choice of local parametrizations, the tensor representation does not! Therefore, we would strongly prefer the tensor representation when we write the first fundamental form.

Exercise 3.7

Show that \(du^i = \dfrac{\partial u_i}{\partial x_k}\,dx^k\). Hint: What is \(du^i\!\left(\pd{}{x_k}\right)\)?

Need help?
Exercise 3.8

Show that the first fundamental form of the graph \(\Sigma_f\) in Example 2.29 is given by: \[ g = \sum_{i=1}^n dx^i\otimes dx^i + df\otimes df, \] where \(df = \dfrac{\partial f}{\partial x_j}\,dx^j\).

Need help?
Exercise 3.9

Compute the first fundamental form of the sphere using the stereographic parametrizations \(F_+\) and \(F_-\) in Example 2.30. Express the form using tensor notations.

Need help?
Remark 3.6

As the first fundamental form \(g\) is a bilinear form that “consumes” two tangent vectors and outputs a scalar, we often call such a quantity to be a 2-tensor.

Checkpoint
Why is the first fundamental form \(g\) called a 2-tensor? (See Remark 3.6.)
WhyA tensor of order 2 is a multilinear map that takes two vector arguments. As noted in Remark 3.6, \(g(X,Y)\) consumes two tangent vectors and outputs a scalar, which is precisely why it earns the name 2-tensor.