When can we ignore the delta function in th Feynman rules?

1. May 29, 2008

Vereinsamt

in peskin-schroeder and here the amplitude for $$e^-e^+\rightarrow \mu^- \mu^+$$ is written using feynman rules as follows
$$-iM=[\bar{v}(p_2)(-ie\gamma^\mu )u(p_1)] \frac{-ig_{\mu\nu}}{q^2}[\bar{u}(k_1)(-ie\gamma^\nu )v(k_2)]$$

but what about the delta function integeration? is it already done here?