Chapter 2 › 2.4 Maps and Functions from Surfaces › 2.4.1 Tangent Maps between Regular Surfaces
Subsection

2.4.1  Tangent Maps between Regular Surfaces

Given a smooth map \(\Phi:M\to N\) between two regular surfaces \(M\) and \(N\), there is a naturally defined map called the tangent map, denoted by \(\Phi_*\) in this course, between the tangent planes \(T_pM\) and \(T_{\Phi(p)}N\).

Let us consider a smooth local parametrization \(\bF(u_1,u_2):U_M\to M\). The composition \(\Phi\circ\bF\) can be regarded as a map from \(U_M\) to \(\R^3\), so one can talk about its partial derivatives \(\pd{(\Phi\circ\bF)}{u_i}\): \[\pd{\Phi}{u_i}(\Phi(p)):=\pd{(\Phi\circ\bF)}{u_i}\Big|_{(u_1,u_2)}=\frac{d}{dt}\Big|_{t=0}\Phi\circ\bF\big((u_1,u_2)+t\,\hat{\mathbf{e}}_i\big)\] where \((u_1,u_2)\) is a point in \(U_M\) such that \(\bF(u_1,u_2)=p\). The curve \(\bF((u_1,u_2)+t\,\hat{\mathbf{e}}_i)\) is a curve on \(M\) with parameter \(t\) along the \(u_i\)-direction. The curve \(\Phi\circ\bF((u_1,u_2)+t\,\hat{\mathbf{e}}_i)\) is then the image of the \(u_i\)-curve of \(M\) under the map \(\Phi\) (see Figure 2.11). It is a curve on \(N\), so \(\pd{\Phi}{u_i}\) is a tangent vector to the surface \(N\).

<b>Figure 2.11.</b> Partial derivative of the map \(\Phi:M\to N\).
Figure 2.11. Partial derivative of the map \(\Phi:M\to N\).
Exercise 2.26

Denote \(S^2\) to be the unit sphere \(x^2+y^2+z^2=1\). Let \(\Phi:S^2\to S^2\) be the rotation map about the \(z\)-axis defined by \(\Phi(x,y,z)=(x\cos\alpha-y\sin\alpha,\ x\sin\alpha+y\cos\alpha,\ z)\) where \(\alpha\) is a fixed angle. Calculate the following partial derivatives under the given local parametrizations:

  1. \(\pd{\Phi}{\theta}\) and \(\pd{\Phi}{\varphi}\) under \(\bF(\theta,\varphi)=(\sin\varphi\cos\theta,\sin\varphi\sin\theta,\cos\varphi)\);
  2. \(\pd{\Phi}{u}\) and \(\pd{\Phi}{v}\) under \(\bF_2\) in Example 2.4;
  3. \(\pd{\Phi}{u}\) and \(\pd{\Phi}{v}\) under \(\bF_+\) in Example 2.5.
Need help?

Next, we write the partial derivative \(\pd{\Phi}{u_i}\) in a fancy way. Define \(\Phi_*\big(\pdF{u_i}\big):=\pd{\Phi}{u_i}.\) Then, one can regard \(\Phi_*\) as a map that takes the tangent vector \(\pdF{u_i}\) in \(T_pM\) to another vector \(\pd{\Phi}{u_i}\) in \(T_{\Phi(p)}N\). Since \(\big\{\pdF{u_i}(p)\big\}\) is a basis of \(T_pM\), one can then extend \(\Phi_*\) linearly and define it as the tangent map of \(\Phi\). Precisely, we have:

Definition 2.24 (Tangent Maps)

Let \(\Phi:M\to N\) be a smooth map between two regular surfaces \(M\) and \(N\) in \(\R^3\). Let \(\bF:U_M\to M\) and \(\bG:U_N\to N\) be two smooth local parametrizations covering \(p\) and \(\Phi(p)\) respectively. Then, the tangent map of \(\Phi\) at \(p\in M\) is denoted by \((\Phi_*)_p\) and is defined as: \[(\Phi_*)_p:T_pM\to T_{\Phi(p)}N,\qquad(\Phi_*)_p\left(\sum_{i=1}^{2}a_i\,\pdF{u_i}(p)\right)=\sum_{i=1}^{2}a_i\,\pd{\Phi}{u_i}(\Phi(p)).\] If the point \(p\) is clear from the context, \((\Phi_*)_p\) can be simply denoted by \(\Phi_*\).

Checkpoint
Definition 2.24 defines \((\Phi_*)_p:T_pM\to T_{\Phi(p)}N\) by \((\Phi_*)_p\big(\sum a_i\pdF{u_i}\big)=\sum a_i\pd{\Phi}{u_i}\). What kind of map is \((\Phi_*)_p\)?
WhyThe formula respects scaling and addition of the coefficients \(a_i\), so \((\Phi_*)_p\) is a linear transformation \(T_pM\to T_{\Phi(p)}N\) — the differential of \(\Phi\) at \(p\).
Remark 2.25

Some textbooks may use \(d\Phi_p\) to denote the tangent map of \(\Phi\) at \(p\). \(\square\)

Checkpoint
Remark 2.25 notes some textbooks denote the tangent map of \(\Phi\) at \(p\) by which alternative symbol?
WhyThe tangent map \((\Phi_*)_p\) is commonly written \(d\Phi_p\) (the differential) in many references; the two notations mean the same linear map between tangent planes.
Example 2.26

Consider the unit sphere \(S^2\) locally parametrized by \(\bF(\theta,\varphi)=(\sin\varphi\cos\theta,\sin\varphi\sin\theta,\cos\varphi)\) and the rotation map \(\Phi(x,y,z)=(x\cos\alpha-y\sin\alpha,x\sin\alpha+y\cos\alpha,z)\). From Exercise 2.26, one should have figured out that \[\begin{aligned}\pd{\Phi}{\theta}&=(-\sin\varphi\sin(\theta+\alpha),\ \sin\varphi\cos(\theta+\alpha),\ 0),\\[4pt]\pd{\Phi}{\varphi}&=(\cos\varphi\cos(\theta+\alpha),\ \cos\varphi\sin(\theta+\alpha),\ -\sin\varphi).\end{aligned}\]

Next we want to write them in terms of the basis \(\big\{\pd{\bF}{\theta},\pd{\bF}{\varphi}\big\}\). However, we should be careful about the base points of these vectors. Consider a point \(p\in S^2\) with local coordinates \((\theta,\varphi)\); the vectors \(\pd{\Phi}{\theta}\) and \(\pd{\Phi}{\varphi}\) computed above are based at the point \(\Phi(p)\) with local coordinates \((\theta+\alpha,\varphi)\). Therefore, we should express them in terms of the basis \(\big\{\pd{\bF}{\theta}(\Phi(p)),\pd{\bF}{\varphi}(\Phi(p))\big\}\), not \(\big\{\pd{\bF}{\theta}(p),\pd{\bF}{\varphi}(p)\big\}\). At \(\Phi(p)\), we have \[\pd{\bF}{\theta}(\Phi(p))=(-\sin\varphi\sin(\theta+\alpha),\sin\varphi\cos(\theta+\alpha),0)=\pd{\Phi}{\theta}(\Phi(p)),\] \[\pd{\bF}{\varphi}(\Phi(p))=(\cos\varphi\cos(\theta+\alpha),\cos\varphi\sin(\theta+\alpha),-\sin\varphi)=\pd{\Phi}{\varphi}(\Phi(p)).\]

Therefore, the tangent map \((\Phi_*)_p\) acts on the basis vectors by \((\Phi_*)_p\big(\pd{\bF}{\theta}(p)\big)=\pd{\bF}{\theta}(\Phi(p))\) and \((\Phi_*)_p\big(\pd{\bF}{\varphi}(p)\big)=\pd{\bF}{\varphi}(\Phi(p))\). In other words, the matrix representation \([(\Phi_*)_p]\) with respect to the bases \(\big\{\pd{\bF}{\theta}(p),\pd{\bF}{\varphi}(p)\big\}\) for \(T_pS^2\) and \(\big\{\pd{\bF}{\theta}(\Phi(p)),\pd{\bF}{\varphi}(\Phi(p))\big\}\) for \(T_{\Phi(p)}S^2\) is the identity matrix. However, it is not perfectly correct to say \((\Phi_*)_p\) is an identity map, since the domain and co-domain are different tangent planes. \(\square\)

Checkpoint
In Example 2.26 the rotation acts on the sphere chart by \(\bF^{-1}\circ\Phi\circ\bF:(\theta,\varphi)\mapsto(\theta+\alpha,\varphi)\). What subtlety does the example stress when writing \(\pd{\Phi}{\theta},\pd{\Phi}{\varphi}\) in the basis \(\{\pd{\bF}{\theta},\pd{\bF}{\varphi}\}\)?
WhyThe tangent map sends vectors at \(p\) to vectors at \(\Phi(p)\). Since \(\Phi\) shifts \(\theta\) by \(\alpha\), the basis at the image point is \(\{\pd{\bF}{\theta},\pd{\bF}{\varphi}\}\) evaluated at \(\Phi(p)\); care with base points is essential.
<b>Figure 2.12.</b> Tangent maps of rotations.
Figure 2.12. Tangent maps of rotations.
Exercise 2.27

Let \(\Phi\) be as in Example 2.26. Consider the stereographic parametrization \(\bF_+(u,v)\) defined in Example 2.5. Suppose \(p\in S^2\), express the matrix representation \([(\Phi_*)_p]\) with respect to the bases \(\big\{\pd{\bF_+}{u},\pd{\bF_+}{v}\big\}_p\) and \(\big\{\pd{\bF_+}{u},\pd{\bF_+}{v}\big\}_{\Phi(p)}\).

Need help?