while reviewing the book on differential geometry by Manfredo do Carmo, I came across the equation of lines of curvature. Do Carmo roughly states the following: A connected regular curve $C$ in a coordinate neighborhood $x$ is a line of curvature if, and only if for any parametrization $\alpha(t) = x(u(t), v(t))$, $t \in I$, of $C$, the following equation is satisfied: $$ N(u,v)(v^{\prime})^{2} + M(u,v) u^{\prime} v^{\prime} + L(u,v) (u^{\prime})^{2} = 0,$$ where $N(u,v)= gF - fG, M(u,v)= gE - eG, N(u,v)= fE - eF$, and $E,F,G$ are coefficients of the first fundamental form, and $e,f,g$ are coefficients of the second fundamental form.

My questions are as follows:

  1. If I solve the equation, do I find "directions" in the coordinate plane where, whenever I take derivatives of curves in those directions, their tangent vectors at the derived point are the greatest or smallest? (As, for example, in the case of surfaces of revolution, where I would have $x(u_{0},v)$ or $x(u,v_{0})$).

  2. If I find the points $(u_{0},v_{0})$ where the functions $N=M=L=0$, am I finding umbilical points? If not, what do I find through this process?

Thank you in advance.

  • $\begingroup$ Directions ( in Euler relation ) associated with max/min normal curvatures. $\endgroup$
    – Narasimham
    Jul 29, 2023 at 20:14
  • $\begingroup$ Hi @Narasimham. I think the same in the first question, but, in the second question? $\endgroup$
    – Skinner.
    Jul 30, 2023 at 15:16
  • $\begingroup$ Yes, $\mathrm{II} = c\mathrm I$ iff the shape operator is $c$ times the identity map. $\endgroup$ Jul 30, 2023 at 21:51
  • $\begingroup$ Hi, @TedShifrin I don't understand your answer, can you explain me? $\endgroup$
    – Skinner.
    Jul 31, 2023 at 13:45

1 Answer 1


No, we have proportionality of coefficients in the two fundamental forms. We have in the alternate notation of normal curvature. Let


$$\frac{E}{L}=\frac{F}{M}=\frac{G}{N} = \mu, \text{a constant that need not be necessarily positive.}$$

It can be shown for surface $r$ $$N_1= -\mu \cdot r $$

Differentiate w.r.t $v$ and $u$ and subtracting,

$$ N_{12}= N_{21}, ~\mu_1r_2=\mu_2 r_1;$$ since

$$ r_1 \text{ is not parallel to } r_2, \mu s \text { are constant making }$$

$$ \mu_1=0,~\mu_2=0 $$

If constant $\mu$ is non-zero

$$r_1=-N_1/\mu, \text{ it integrates to } r=-N_1/\mu+r_0$$

with $r_o$ as a constant of integration

$$ (r-r_o)^2 =\frac{N^2}{\mu^2}~=\frac{1}{\mu^2}~$$ Thus $r$ locus is a sphere.

If $\mu$ is zero, $N_1=N_2=0,$ the $N$ vector itself is constant and the surface is a plane.

The normal curvature is independent of curvature ratios $(du,dv), $ same value holding good for all directions, so all such points are umbilics on the surface.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .