This question originates from proof of Proposition 8.32 of the John Lee's Introduction to Riemannian Manifolds book. It seems easy calculation but I don't understand more rigorously.

Let $(M,g)$ be a Riemannian $n$-manifold and $p\in M$. Let $v \in T_pM$ be a unit vector. Let $(b_1, \dots , b_n)$ be any orthonormal basis for $T_pM$ with $b_1 =v$. Then, why $$ Rc_p(v,v) = R_{11}(p) = R^{k}_{k11}(p) = \Sigma_{k=1}^{n}Rm_p(b_k, b_1, b_1, b_k) $$

? I think that I am unfamiler to the definition of Ricci ( Riemannian ) curvature tensor-and its relation to component- Can any one give more detailed explanation?

  • $\begingroup$ What then is your definition of the Ricci tensor? $\endgroup$ Commented Jun 25, 2023 at 5:29
  • $\begingroup$ Uhm.. for vector fields $X,Y$, $Rc(X,Y) = \operatorname{tr}(Z \mapsto R(Z,X)Y)$. And I don't know how to cennect this abstract definition and the concrete calculation involving component of $Rc$ above. $\endgroup$
    – Plantation
    Commented Jun 25, 2023 at 5:44
  • 1
    $\begingroup$ Take $X=Y=v=b_1$ and take $Z=b_k$, $k=1,\dots,n$. You still need to convert from the $(3,1)$-tensor definition you just gave me of curvature to the $(4,0)$-tensor form, which Lee is using here. $\endgroup$ Commented Jun 25, 2023 at 5:59
  • $\begingroup$ Thank you.. And I still don't get it completely :) I don't know how to deal with the trace in the definition of $Rc$. I think that I am missing some key point. $\endgroup$
    – Plantation
    Commented Jun 25, 2023 at 6:19
  • 1
    $\begingroup$ Summing on $k$ in $\sum R^k_{k11}$ is the trace. $\endgroup$ Commented Jun 25, 2023 at 6:26

1 Answer 1


Just an additional comment:

Say, $\mathbf{T}$ is an arbitrary $(1,3)$-tensor. Its components are by definion the functions on $M$ which allow to express the tensor as linear combination $$ \mathbf{T}={T^i}_{j\,k\,\ell}\;\partial_i\otimes dx^j\otimes dx^k\otimes dx^\ell\ $$ (use summation over all indices).

From the fact that basis vector fields $\partial_i$ and basis one-forms $dx^j$ are dual, $$\partial_i(dx^j)={\delta_i}^j\,,\quad dx^j(\partial_i)= {\delta^j}_i\,,$$ it is obvious how to get the components: let $\mathbf T$ eat $dx^i,\partial_j,\partial_k,\partial_\ell$ and you have ${T^i}_{j\,k\,\ell}\,.$

For simplicity I explained this in a coordinate basis but it works with your basis $b_1,\dots,b_n$ instead of $\partial_1,\dots,\partial_n$ as well. Just take the corresponding basis that is dual to $b_1,\dots,b_n\,.$ Of course the components in this basis are different but they are extracted in exactly the same way.


You must log in to answer this question.

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