In "Quantum Gravity in 2+1 dimension" by S Carlip, Sec 3.1 (where the metric of a spacetime with a point source is derived, using the ADM formalism), equation 3.8 states that (this is the momentum constraint for a metric on a constant surface),

$$ \nabla_{j} \Pi^{ij} = 0 = g^{il}\partial_{k} \Pi^{k}_{l} - (1/2)g^{il}(\partial_{l}g_{jk})\Pi^{jk}. $$

There should be two terms with Christoffel connections from the covariant derivative, however this has only one. I am getting another term which I find to be non-zero. Can somebody please tell me how the other term vanishes?


1 Answer 1


Recall that momentum $\pi^{ij}$ is a tensor density in 2-space, cf. e.g. Ref. 2. In other words, $$\frac{\pi^{ij}}{\sqrt{\det g}}$$ is a symmetric (2,0) tensor. Therefore the covariant derivative is

$$(\nabla_{\ell}\pi)^{ij}~=~ \partial_{\ell} \pi^{ij} + \Gamma^i_{\ell k} \pi^{kj} + \pi^{ik}\Gamma^j_{\ell k} -\pi^{ij} \partial_{\ell}\ln\sqrt{\det g}.$$

The contraction becomes

$$(\nabla_j\pi)^{ij}~=~ \partial_j \pi^{ij} + \Gamma^i_{jk} \pi^{kj} + \pi^{ik}\underbrace{\left(\Gamma^j_{jk} - \partial_k\ln\sqrt{\det g}\right)}_{=0}$$ $$~=~\partial_j \pi^{ij} + g^{i\ell} \left(\partial_{j} g_{\ell k}-\frac{1}{2}\partial_{\ell} g_{jk}\right)\pi^{kj} $$ $$~=~g^{i\ell} \left( \partial_{j} \left(g_{\ell k}\pi^{kj}\right) -\frac{1}{2}\pi^{kj}\partial_{\ell} g_{jk}\right), \tag{3.8} $$ in agreement with eq. (3.8) in Ref. 1.


  1. S. Carlip, Quantum Gravity in 2+1 Dimensions, Cambridge University Press, 1998, Section 3.1, p. 39.

  2. C.W. Misner, K.S. Thorne, and J.A. Wheeler, Gravitation, 1973, Section 21.7, p. 521.

  • $\begingroup$ Thanks a lot! I was treating $\pi^{ij}$ as a tensor instead of a tensor density and so I was getting wrong results. $\endgroup$
    – Sourav
    Commented Mar 23, 2014 at 19:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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