2

In my class, we defined the tangent space $T_pS$ of a regular curve $S$ at a point $p\in S$ as the set $\gamma'(0)$ such that $\gamma$ is a regular curve in $S$ and $\gamma(0)=p$.

My particular problem involves a torus $S$ parametrized as $\sigma(\theta,t)=((2+\cos{t})\cos{\theta},(2+\cos{t})\sin{\theta},\sin{t})$, and we want to describe the tangent plane at an arbitrary point of the $t=0$ parameter.


Solution attempt

Since we already have a parametrized surface, I assume we can treat $\sigma$ as a coordinate chart $U\to S$ for some $U\subset\mathbb{R}^2$. Let $p=\sigma(\theta,0)\in S$. We know that $\sigma$ is a diffeomorphism, so $q=\sigma^{-1}(p)=(\theta,0)$. Then $T_pS$ is spanned by the vectors $\sigma_\theta(q)$ and $\sigma_t(q)$. Now we compute these vectors...

$\sigma_\theta(\theta,0)=(-3\sin\theta,3\cos\theta,0)$

$\sigma_t(\theta,0)=(0,0,1)$

I want to believe these computations are correct because the span will give us a two-dimensional subspace of $\mathbb{R}^3$, which is in-line with a lemma stated in our text about the dimension of $T_pS$. But how exactly would we "describe" this span? Visually speaking, I know the first partial derivative is giving us a circle with radius 3 whereas the second one gives us a fixed point. But it feels like I am simply getting my torus back when I consider linear combinations of these two vectors.

  • 1
    You shouldn't think of the first partial derivative giving you a circle. For each choice of point $(\theta, 0)$ on the torus, you get vectors spanning the tangent space at that point. For example, with $\theta = 0$, your two vectors are $(0,3,0)$ and $(0,0,1)$ and these span the tangent space at the point $(0,0)$. Different point, different pair of vectors. (Note that an unfortunate parametrization could make these (surprise!) parallel at one or more points (so you would need to find a second vector another way), but this doesn't happen in this problem.) – Eric Towers Feb 24 '22 at 03:49
  • 1
    Ah, I'm starting to see my mistake now. $p$ is fixed here, so it makes little sense to think of $\sigma_{\theta}$ as a circle. I think I'll be able to solve this problem now. – hello4649 Feb 24 '22 at 03:54

0 Answers0