I try to understand the difference between the classical definition of covariant derivative :


(where index $i$ represents curvilinear cooordinates $x^{i}$ and $V^{j}$ the $j-\text{th}$ component of vector $V$)

and the definition of called " covariant derivative of a vector field $V$ along a vector field $Z$ " and noted :


What's the expression of (2) and how to make the link with equation (1) ?

I saw that equation (2) was used like this in the demonstration of Ricci theorem :

$$Z \langle X,Y \rangle = \langle \nabla_Z X, Y\rangle + \langle X, \nabla_Z Y\rangle$$

Sorry if this question seems to be evident but I am just starting to learn basics of geometry differential.

Thanks for your help

  • $\begingroup$ In short, the link is that $\nabla_{\frac{\partial}{\partial x^i}} V = \sum_{j=1}^n \nabla_i V^j \frac{\partial }{\partial x^j}$, $\endgroup$
    – user99914
    Feb 5, 2017 at 8:07
  • $\begingroup$ :John Ma, thanks. So vector field Z would be represented by $\dfrac{\partial}{\partial x^{i}}$ ? Is this term representing also the definition of local basis vectors, i.e $\vec{e_{i}}=\dfrac{\partial \vec{OM}}{\partial x^{i}}$ (with $\vec{OM}$ the position vector )? In this case, could we write $\nabla_{\vec{e_{i}}}\vec{V}=\nabla_{i}V^{j}\,\vec{e_{j}}=\nabla_{i}(\vec{V}\cdot\vec{e^{j}}) \vec{e_{j}}= \nabla_{i}\vec{V}$? It seems I do confusions $\endgroup$
    – youpilat13
    Feb 5, 2017 at 9:01

1 Answer 1


After choosing coordinates $(x^1,\dots,x^n)$, each vector field $V$ can be expressed locally as $V = V^j \frac{\partial}{\partial x^j}$. If you are given a covariant derivative operator $\nabla$, the Christoffel symbols with respect to the coordinate system $(x^1,\dots,x^n)$ are defined by the equation

$$ \nabla_{\frac{\partial}{\partial x^i}} \frac{\partial}{\partial x^j} = \Gamma_{ij}^k \frac{\partial}{\partial x^k}. $$

Namely, $\Gamma_{ij}^k$ gives you the $\frac{\partial}{\partial x^k}$-th component of the covariant derivative of $\frac{\partial}{\partial x^j}$ in the direction $\frac{\partial}{\partial x^i}$. Using the product rule of the covariant derivative, we have

$$ \nabla_{\frac{\partial}{\partial x^i}} \left( V^j \frac{\partial}{\partial x^j} \right) = \frac{\partial V^j}{\partial x^i} \frac{\partial}{\partial x^j} + V^j \nabla_{\frac{\partial}{\partial x^i}} \left( \frac{\partial}{\partial x^j} \right) = \frac{\partial V^j}{\partial x^i} \frac{\partial}{\partial x^j} + V^j \Gamma_{ij}^k \frac{\partial}{\partial x^k}. $$

This can be abbreviated to

$$ (\nabla_i V)^k = \partial_i(V^k) + V^j \Gamma_{ij}^k $$

(so the $k$-th component of the covariant derivative $\nabla_i V$ is given by the regular derivative $\partial_i(V^k)$ of the $k$-th component plus a "correction factor" which involves the all the components of $V$ and the Christoffel symbols).

  • $\begingroup$ :levap thanks. How do you prove that each vector field $V$ can be expressed locally as $V=V^{j}\,\dfrac{\partial}{\partial x^{j}}$ ? $\endgroup$
    – youpilat13
    Feb 8, 2017 at 10:59
  • $\begingroup$ You can't differentiate a vector field with respect to a vector field in a coordinate independent way without introducing a connection and using the covariant derivative. $\endgroup$
    – levap
    Feb 8, 2017 at 16:09
  • $\begingroup$ do you agree that I can also write : $\nabla_{\frac{\partial}{\partial x^i}} \dfrac{\partial}{\partial x^j} =\dfrac{\partial^{2}}{\partial x^{i}\partial x^{j}} =\Gamma_{ij}^k \dfrac{\partial}{\partial x^k}$ ? $\endgroup$
    – youpilat13
    Feb 8, 2017 at 18:33
  • $\begingroup$ You can write whatever you want but that is a very bad choice of notation. $\endgroup$
    – levap
    Feb 8, 2017 at 19:07
  • $\begingroup$ :levap thanks. And what about the proof of the equation that locally, vector field $V$ is : $V=V^{j}\,\dfrac{\partial}{\partial x^{j}}$ ? I don't know how you get this equality. regards $\endgroup$
    – youpilat13
    Feb 8, 2017 at 20:12

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.