MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $M$ be a pseudo-Riemannian manifold. Assume that $M$ comes with a zero-torsion affine connexion $\nabla.$ There is no need for $\nabla$ to be the Levi-Civita connexion. Recall that the curvature tensor of $\nabla$ is given by $R(X,Y)Z = \nabla_X\nabla_YZ - \nabla_Y\nabla_XZ - \nabla_{[X,Y]}Z,$ while the Ricci tensor is given by the trace of the curvature tensor: $\mbox{Ric}(Y,Z) = \mbox{trace}\left[ X \mapsto R(X,Y)Z \right].$

The connexion $\nabla$ is called locally equi-affine if around each point $x \in M$ there is a parallel volume form, i.e. a non-vanishing, skew-symmetric, $n$-form such that $\nabla\omega = 0.$

In Nomizu & Sasaki's "Affine Differential Geometry", in Prop 3.1 on page 14, they say that a zero-torsion connexion has symmetric Ricci tensor if and only if it is locally equi-affine.

In fact, there is a differential one-form $\tau$ such that $\nabla_X\omega = \tau(X)\omega$ for any vector field $X \in \mathfrak{X}(M).$ Thus, we see that $\nabla$ has symmetric Ricci tensor if and only if $\tau \equiv 0.$

However, in Section 4 of page 237, they say that $\nabla$ has symmetric Ricci tensor if and only if the exterior derivative $\mbox{d}\tau \equiv 0.$

Clearly only one of them is correct, but the proof to Prop 3.1. is so sparing with details that I can't follow it. It's one of those proofs that if you can't prove the statement thenyou can't understand the proof.

So, is it $\tau \equiv 0$ or $\mbox{d}\tau \equiv 0$?

share|cite|improve this question
up vote 8 down vote accepted

They are both correct. :-)

I gave a somewhat detailed write-up last year on my blog, but the gist of the argument is that if $\tau = 0$, then $d\tau = 0$. On the flip side, if $d\tau = 0$, locally you can lift $\tau = du$ for some function $u$, and by replacing $\omega$ with $e^{-u}\omega$ you construct the desired equi-affine volume form.

In other words, the second statement applies to arbitrary volume forms, while the first statement that $\tau$ vanishes applies to a special volume form, the one that gives the equi-affine condition.

Just to make clear that there's no contradiction. We first note that for any volume form (non-vanishing differential form of top degree), the expression $\nabla_X \omega = \tau(X)\omega$ holds for some one-form $\tau$ depending on $\omega$. Since we are interested in local statements, assume now our manifold is diffeomorphic to the unit ball.

The first statement says that "$\exists \omega$ a volume form such that $\tau = 0$."

The second statement say that "$\forall \omega$ a volume form, the corresponding $d\tau = 0$."

The two statements are equivalent using what I sketched in the second paragraph above.

share|cite|improve this answer
That's brilliant, thanks. Is there some way for my to communicate with you other than this comment box? It only allows 500 characters, and I'd really like to make sure I understand this properly. It's really refreshing to find someone that knows what I'm on about! – Fly by Night Jun 27 '11 at 21:37

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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