# Variation of gamma and Christoffel

I am trying to experiment with Lagrangian densities and came across a term similar to $$\gamma^i\Gamma^j_{ik}A^k$$ where the $\gamma$ are the gamma matrices, $\Gamma$ are the Christoffel Symbols and $A$ are the field components.

In the variation of the term my question is somewhat twofold: is there any "nice" way to compute $\dfrac{\delta}{\delta g^{lm}}\gamma^a$, knowing the relation $\{\gamma^i,\gamma^j\}=2g_{ij}\mathbf{I}$?

Secondly, from what I have seen $\delta\Gamma^i_{jk}=\Gamma_{jkl}\delta g^{il}+g^{il}\delta\Gamma_{jkl}$, commonly rewritten in a different form. The question Variation of the Christoffel symbols with respect to $g^{\mu\nu}$ does not seem to have a satisfactory answer for my purpose; it seems like the variation terms will stay in the field equation so it does not make sense to leave variations in the expression. Is it valid to naively assume that $\dfrac{\delta}{\delta g^{lm}}\Gamma^j_{ik}=\Gamma_{jkl}$?

• I don't think $\frac{\delta}{\delta g^{lm}}\gamma^a$ is a meaningful expression. $\gamma^a$ doesn't depend on $g^{lm}$, but on $e^a_i$ instead. The object you should be computing is $\frac{\delta}{\delta e^a_i}\gamma^b$ (which is quite simple to evaluate). – AccidentalFourierTransform Jun 19 '18 at 21:29
• @AccidentalFourierTransform, I appreciate the comment but would the variation with respect to the tetrad have any significance in the E-L equations? I'm not sure how it relates considering the equations are for the metric and field. – Quantumness Jun 19 '18 at 21:38