Appropriate specification of neurotransmitters is a key feature of neuronal network assembly. There is much evidence that genetic programs are responsible for this aspect of cell fate and neuronal differentiation. Are there additional ways in which these processes are shaped? Recent findings demonstrate that altering patterned Ca(2+) spike activity that is spontaneously generated by different classes of embryonic spinal neurons in vivo changes expression of neurotransmitters in a homeostatic manner, as if to achieve a constant level of excitation. Activity-dependent changes in presynaptic transmitter expression pose a matching problem: are there corresponding changes in postsynaptic transmitter receptor expression, or are axons rerouted to novel targets with which functional synapses can be formed?